WO2015093457A1 - Information processing device, information processing method, program, and recording medium - Google Patents

Information processing device, information processing method, program, and recording medium Download PDF

Info

Publication number
WO2015093457A1
WO2015093457A1 PCT/JP2014/083198 JP2014083198W WO2015093457A1 WO 2015093457 A1 WO2015093457 A1 WO 2015093457A1 JP 2014083198 W JP2014083198 W JP 2014083198W WO 2015093457 A1 WO2015093457 A1 WO 2015093457A1
Authority
WO
WIPO (PCT)
Prior art keywords
shape data
shape
contour
point
unit
Prior art date
Application number
PCT/JP2014/083198
Other languages
French (fr)
Japanese (ja)
Inventor
五十嵐　健夫
Original Assignee
国立大学法人東京大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 国立大学法人東京大学 filed Critical 国立大学法人東京大学
Publication of WO2015093457A1 publication Critical patent/WO2015093457A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
    • G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
    • G06V10/752—Contour matching

Definitions

  • the present invention relates to an information processing apparatus, an information processing method, a program, and a recording medium that perform matching processing.
  • Non-Patent Document 1 Non-Patent Document 1
  • Non-Patent Documents 2 and 3 there is a technology that performs matching using the distance on the surface between two points of the 3D shape and a technology that performs matching using the Euclidean distance between two points on the surface of the 3D shape.
  • the inner region is often distorted when texture transformation is performed using the contour after matching. This is considered to be caused by performing matching using a contour line or a surface shape point without considering the internal region.
  • An object is to provide a recording medium.
  • An information processing apparatus includes an input unit that inputs first shape data and second shape data, a sampling unit that samples each point in the outline of the first shape data and the second shape data, and For each point sampled in the first shape data, a first shape descriptor based on an internal region of the first shape data, and for each point sampled in the second shape data, the second shape A setting unit for setting a second shape descriptor based on an internal area of the data, comparing the first shape descriptor and the second shape descriptor, and each point in the contour of the first shape data; A matching unit that associates each point in the contour of the second shape data, a point in the contour of the first shape data that is associated, and a contour in the contour of the second shape data And an output unit for outputting a set of.
  • the setting unit as the first shape descriptor, a first geodetic that rearranges the distance through the first shape data between the sampled point and each point in the first shape data
  • a second geodesic distance obtained by generating a distance profile and rearranging the distance passing through the inside of the second shape data between the sampled point and each point inside the second shape data as the second shape descriptor A profile may be generated.
  • a geodesic distance profile suitable as a shape descriptor can be generated.
  • the setting unit may divide the first shape data and the second shape data into a plurality of regions, and use points in each region as internal points. Thereby, the setting of a suitable internal point can be made easy.
  • the setting unit divides the first shape data and the second shape data into a plurality of regions by using a triangular mesh division, and a point in each region may be used as a center point of the divided triangle. Good. Thereby, the correspondence between the region and the internal point can be clarified.
  • the setting unit may generate the first geodesic distance profile and the second geodesic distance profile by rearranging the distances in consideration of the area of each region. Thereby, it is possible to generate a geodesic distance profile that more reflects the characteristics of the internal region.
  • the matching unit calculates a logarithm of the first geodetic distance profile and a logarithm of the second geodetic distance profile, and performs an association process so that a sum of differences between the calculated logarithms is minimized. You may go. Thereby, the matching accuracy can be further improved.
  • the matching unit performs a filtering process on the first geodetic distance profile and the second geodetic distance profile, and performs an association process so that a difference between the geodetic distance profiles after the filtering process is minimized. May be performed. Thereby, the outline matching process according to a use can be performed.
  • the setting unit may determine whether the first shape data is between the first shape data and the first closed region.
  • a third shape descriptor of each point in the contour of the first closed region is set based on the region in the first closed region, and the second shape data and the second closed region are used to determine the third shape descriptor.
  • You may set the 4th shape descriptor of each point in the outline of 2 closed areas.
  • the sampling unit may enlarge or reduce the first shape data or the second shape data so that the internal area of the first shape data matches the internal area of the second shape data. . Thereby, even if the size of the shape data to be matched is different, the contour matching process can be appropriately performed.
  • An information processing method is an information processing method executed by a computer, the step of inputting first shape data and second shape data, the first shape data, and the second shape data. Sampling each point in the contour of the shape data; for each point sampled in the first shape data, a first shape descriptor based on an internal region of the first shape data; and in the second shape data For each sampled point, setting a second shape descriptor based on an internal region of the second shape data; comparing the first shape descriptor and the second shape descriptor; and Associating each point in the contour of the first shape data with each point in the contour of the second shape data, and the contour of the associated first shape data Including a point where definitive, and outputting a set of points in the second shape data profile, a.
  • the predetermined aspect of the present invention it is possible to perform the matching process on the contour of the shape in consideration of the internal region of the shape.
  • FIG. 1 It is a flowchart which shows an example of the production
  • FIG. It is a figure which shows the texture texture-transformed by the outline matching in Experiment 1.
  • FIG. It is a figure which shows an example of the shape data used for Experiment. It is a figure which shows the distortion condition of the internal area
  • FIG. It is a figure which shows the texture texture-transformed by the outline matching in Experiment 2.
  • FIG. It is a figure which shows an example of the shape data used for Experiment 3. It is a figure which shows the experimental result of the experiment 3.
  • FIG. It is a figure which shows the experimental result of the experiment 4.
  • FIG. 1 is a block diagram illustrating an example of a schematic configuration of an information processing apparatus 10 according to the embodiment.
  • the information processing apparatus 10 includes a central processing unit (CPU) 102, a random access memory (RAM) 104, a read only memory (ROM) 106, a drive device 108, a network I / F ( Interface) 110, an input device 112, and a display device 114. These components are connected to each other via a bus so as to be able to transmit and receive data.
  • the CPU 102 is a control unit that controls each device, calculates data, and processes in the computer.
  • the CPU 102 is an arithmetic unit that executes a matching processing program stored in the RAM 104 or the ROM 106.
  • the CPU 102 receives shape data from the input device 112, the network I / F 110, and the like, calculates and processes the data, and outputs the calculation result to the display device 114, the storage device, and the like.
  • the RAM 104 is, for example, a main storage unit.
  • the RAM 104 is a storage device that stores or temporarily stores programs and data such as OS (Operating System) and application software that are basic software executed by the CPU 102.
  • OS Operating System
  • application software that are basic software executed by the CPU 102.
  • the ROM 106 is a storage device that stores data related to application software, for example.
  • the drive device 108 reads the program from the recording medium 116, such as a CD-ROM or SD card, and installs it in the storage device.
  • the recording medium 116 such as a CD-ROM or SD card
  • a predetermined program is stored in the recording medium 116, and the program stored in the recording medium 116 is installed in the information processing apparatus 10 via the drive device 108.
  • the installed predetermined program can be executed by the information processing apparatus 10.
  • the network I / F 110 is an interface between a peripheral device having a communication function and the information processing apparatus 10.
  • the network I / F 110 is connected via a network such as a LAN (Local Area Network) or a WAN (Wide Area Network) constructed by a data transmission path such as a wired and / or wireless line.
  • a network such as a LAN (Local Area Network) or a WAN (Wide Area Network) constructed by a data transmission path such as a wired and / or wireless line.
  • the input device 112 includes a keyboard having cursor keys, numeric input, various function keys, and the like, a mouse and a slide pad for selecting keys on the display screen of the display device 114, and the like.
  • the input device 112 is a user interface for a user to give an operation instruction to the CPU 102 or input data.
  • the display device 114 is configured by an LCD (Liquid Crystal Display) or the like, and performs display according to display data input from the CPU 102. Note that the input device 112 and the display device 114 may be provided outside the information processing apparatus 10.
  • FIG. 2 is a block diagram illustrating an example of functions of the information processing apparatus 10 according to the embodiment.
  • the information processing apparatus 10 illustrated in FIG. 2 includes at least an input unit 202, a sampling unit 204, a setting unit 206, a matching unit 208, and an output unit 210.
  • the CPU 102 can execute the function of each unit by executing a matching processing program.
  • the input unit 202 inputs source shape data and target shape data from the RAM 104, the ROM 106, the network I / F 110, or the input device 112.
  • the source shape data is also referred to as first shape data
  • the target shape data is also referred to as second shape data.
  • the input unit 202 outputs the input first shape data and second shape data to the sampling unit 204.
  • the input unit 202 may input the first shape data serving as the source first and input the second shape data serving as the target later.
  • the sampling unit 204 samples each point in the outline of the first shape data and the second shape data.
  • the sampling unit 204 samples a plurality of points in the contour of each shape data.
  • the sampling unit 204 samples the contour points so as to be equally spaced in the contours of the first shape data and the second shape data, for example.
  • a contour point refers to a point on the contour.
  • the sampling unit 204 outputs the contour points sampled in the first shape data and the contour points sampled in the second shape data to the matching unit 208.
  • the setting unit 206 uses the first shape descriptor based on the internal region of the first shape data and the second shape for each point sampled in the second shape data.
  • a second shape descriptor based on the internal area of the data is set.
  • the shape descriptor based on the internal region refers to, for example, a set of distances between a contour point and a plurality of points in the shape data.
  • the setting unit 206 for example, provides a plurality of points (hereinafter also referred to as internal points) in the internal region of the shape data, and rearranges the distances between the sampled contour points and the internal points (geodesic). Let distance profile be a shape descriptor. At this time, when determining the distance between the contour point and the internal point, the setting unit 206 determines the distance passing through the inside of the shape, not the linear distance.
  • the setting unit 206 uses the first geodesic distance profile for the first shape data as the first shape description and the second geodesic distance profile for the second shape data as the second shape descriptor.
  • the setting unit 206 outputs the set first shape descriptor and second shape descriptor to the matching unit 208.
  • the matching unit 208 compares the first shape descriptor and the second shape descriptor, and performs processing for associating each point in the contour of the first shape data with each point in the contour of the second shape data.
  • the matching process includes such an association process.
  • the matching unit 208 for example, the contour point of the first shape data and the contour point of the second shape data so that the difference between the first shape descriptor acquired from the setting unit 206 and the second shape descriptor is minimized. Matching process is performed.
  • the matching unit 208 may obtain the contour point of the second shape data that matches the contour point of the first shape data by using dynamic programming.
  • the matching unit 208 sets the matched contour point of the first shape data and the contour point of the second shape data, and outputs the set to the output unit 210.
  • the contour point is, for example, coordinate data.
  • the output unit 210 outputs a set of matched contour points of the first shape data and contour points of the second shape data.
  • the output unit 210 sets, for example, the contour points of the second shape data matched with each of the contour points of the first shape data, and outputs the set to the subsequent processing unit, the RAM 204, or the ROM 206.
  • FIG. 3 is a diagram illustrating an example of each storage unit included in the RAM 104 according to the embodiment. 3 includes a shape data storage unit 302, a profile storage unit 304, and a result storage unit 306.
  • the shape data storage unit 302 stores the first shape data and the second shape data input by the input unit 202.
  • the shape data storage unit 302 may store a first image having first shape data and a second image having second shape data.
  • the profile storage unit 304 stores a first geodesic distance profile for each contour point of the first shape data and a second geodesic distance profile for each contour point of the second shape data. Details of the geodesic distance profile will be described later.
  • the result storage unit 306 stores the result of the matching process.
  • the result of the matching process is a set of the contour points of the second shape data that have been matched for each contour point of the first shape data.
  • the information processing apparatus 10 that performs the processing described above associates a geodesic distance profile with each contour point, and performs contour point matching processing using the geodesic distance profile, thereby preventing distortion of the internal region. Matching processing can be performed.
  • FIG. 4 is a block diagram illustrating an example of the configuration of the sampling unit 204 in the embodiment.
  • the sampling unit 204 illustrated in FIG. 4 includes an enlargement / reduction unit 402.
  • the enlargement / reduction unit 402 enlarges or reduces the first shape data or the second shape data so that the internal area of the first shape data matches the internal area of the second shape data.
  • the enlargement / reduction unit 402 calculates, for example, the internal areas of the first shape data and the second shape data, respectively, and enlarges or reduces the area of the second shape data so as to match the area of the first shape data.
  • the sampling unit 204 samples contour points for the enlarged or reduced second shape data.
  • the enlargement / reduction unit 402 has been described as a part of the sampling unit 204, but may be provided between the input unit 202 and the sampling unit 204 as a function different from the sampling unit 204.
  • the information processing apparatus 10 can appropriately perform contour matching processing even when the size of the shape data to be matched is different. Further, the reverse process of the process performed by the enlargement / reduction unit 402 may be executed at an appropriate timing if necessary in the subsequent process.
  • FIG. 5 is a block diagram illustrating an example of the configuration of the setting unit 206 in the embodiment.
  • the setting unit 206 includes a profile generation unit 502, and the profile generation unit 502 includes an area division unit 504, a distance calculation unit 506, an area calculation unit 508, and a rearrangement unit 510.
  • the profile generation unit 502 generates a geodesic distance profile by rearranging the distances through the inside of the shape between the sampled contour point and each internal point of the shape data as a shape descriptor by performing the following processing. .
  • the area dividing unit 504 divides the acquired shape data into a plurality of areas. For example, the region dividing unit 504 divides the shape data into triangular meshes using triangular mesh division. The area dividing unit 504 makes it easy to calculate the area and distance of an area described later by using triangular mesh division. The area dividing unit 504 outputs information on the divided areas to the distance calculating unit 506 and the area calculating unit 508.
  • the distance calculation unit 506 calculates the distance to each internal point for each contour point of the first shape data.
  • the distance calculation unit 506 calculates the distance passing through the inside of the shape in order to consider the internal region when determining the distance.
  • the distance calculation unit 506 calculates the distance to each internal point even in the contour point of the second shape data. A method of calculating the distance when using the triangular mesh division will be described later with reference to FIG.
  • the distance calculation unit 506 outputs each distance to each internal point for each contour point calculated in the first shape data and the second shape data to the rearrangement unit 510.
  • the area calculation unit 508 calculates the area of each area of the acquired shape data. For example, when the shape data is divided into triangle meshes, the area calculation unit 508 calculates the area of each triangle in the shape data. The area calculation unit 508 outputs the area of each region in the first shape data and the area of each region in the second shape data to the rearrangement unit 510.
  • the rearrangement unit 510 rearranges the distances calculated at the contour points according to a predetermined standard. For example, the rearrangement unit 510 rearranges the distances in ascending (short) order, and generates a list with the vertical axis as the distance and the horizontal axis as the internal score. This list becomes a geodesic distance profile.
  • the rearrangement unit 510 adjusts the width of the horizontal axis in proportion to the area of the internal point for this list.
  • the horizontal axis represents the internal area of the shape data.
  • FIG. 6 is a diagram for explaining a geodesic distance profile in the embodiment.
  • FIG. 6A is a diagram illustrating an example of shape data and contour points. Points a11, a12, a13, and a14 shown in FIG. 6A represent a plurality of contour points sampled by the sampling unit 204 in the contour of the shape data sd11. A geodesic distance profile is set at each contour point of the shape data sd11 shown in FIG. FIG. 6A does not show a plurality of internal points provided when setting the geodesic distance profile.
  • the geodesic distance profile at the contour point a11 is a set in which the distances passing through the shape data sd11 from the contour point a11 to the respective internal points are rearranged.
  • the geodesic distance profile is preferably generated in consideration of the area of each region inside the shape data sd11.
  • FIG. 6B is a diagram showing an example of a geodesic distance profile in consideration of the area of each region.
  • the geodesic distance profile shown in FIG. 6B is rearranged in order of increasing distance.
  • a profile pr11 illustrated in FIG. 6B is a geodesic distance profile of the contour point a14 illustrated in FIG. 6A
  • a profile pr12 illustrated in FIG. 6B is a contour point a12 illustrated in FIG. 6A.
  • the profile pr13 shown in FIG. 6B is a geodesic distance profile of the contour point a11 shown in FIG. 6A
  • the profile pr14 shown in FIG. It is a geodesic distance profile of the outline point a13 shown to (A).
  • this profile will be described using the geodesic distance profile shown in FIG.
  • a large distance value indicates that the contour point is convex.
  • the distance value is small, it indicates that the contour point is concave. If the contour point is convex, the inner point around the contour point is far from the contour point. If the contour point is concave, the inner point around the contour point is This is due to being close to the contour point.
  • the contour point is located away from the center (or the center of gravity, etc.) of the shape data.
  • the distance value is small in Global, it indicates that the contour point is close to the center of the shape data. This is due to the fact that the closer the contour point is to the center, the smaller the distance from the contour point to the inner point that is farthest from the center as compared to the case where the contour point is far from the center.
  • the geodesic distance profile described above becomes a shape descriptor of the contour point considering the internal region of the shape data.
  • this embodiment by performing contour point matching using this geodesic distance profile, it is possible to perform matching processing in consideration of an internal region that has not been considered in the prior art.
  • the profile generation unit 502 does not need to provide the region division unit 504 and the area calculation unit 508 as long as a plurality of internal points can be appropriately distributed and set in the internal region of the shape data. Therefore, the rearrangement unit 510 may not perform the process of adjusting the list according to the area of each region.
  • the profile generation unit 502 when a closed region exists in the shape data, the profile generation unit 502 generates an internal point in the internal region between the shape data and the closed region when generating a geodesic distance profile at the contour point of the closed region. Is provided. As a result, appropriate contour matching can be performed on the closed region inside the shape data as shown in FIG.
  • FIG. 7 is a block diagram illustrating an example of the configuration of the matching unit 208 in the embodiment.
  • the matching unit 208 shown in FIG. 7 includes a logarithmic calculation unit 602 and a filter processing unit 604. Note that the logarithmic calculation unit 602 and the filter processing unit 604 may be provided as functions different from the matching unit 208.
  • the logarithm calculation unit 602 calculates a logarithm for each distance of the geodesic distance profile. Based on the difference between each distance for which the logarithm is calculated in the first shape data and each distance for which the logarithm is calculated in the second shape data, the matching unit 208 calculates the contour points of the first shape data and the second shape Performs matching processing with contour points of data.
  • the matching unit 208 By calculating the logarithm, when the matching unit 208 performs the matching process, the influence of the difference when the distance is long can be reduced compared to the difference when the distance is short. That is, the matching unit 208 can obtain a more appropriate matching result by performing the matching process using the logarithm calculated distance.
  • the filter processing unit 604 executes filter processing for each distance included in the geodesic distance profile.
  • the filter processing unit 604 uses, for example, a low pass filter or a high pass filter.
  • the matching unit 208 can perform the matching process suitable for increasing the influence of Local and respecting the corners of the shape. Since Local is a part that is sensitive to human vision, the matching unit 208 can perform more intuitive matching processing. Further, the matching unit 208 can apply a high-pass filter to perform matching processing suitable for increasing the influence of Global and suppressing the influence of noise.
  • the filter processing unit 604 may make the user appropriately select whether to use a low-pass filter or a high-pass filter depending on the application.
  • FIG. 8 is a diagram showing an example of shape data. It is assumed that the profile generation unit 502 receives the shape data sd21 shown in FIG.
  • FIG. 9 is a diagram illustrating an example of a plurality of internal points. As illustrated in FIG. 9, the profile generation unit 502 sets a plurality of internal points in the shape data sd21 according to a predetermined algorithm. At this time, the profile generation unit 502 may set the internal points so that they are not as close as possible.
  • FIG. 10 is a diagram illustrating an example of the distance between the contour point and the internal point.
  • the distance calculation unit 506 includes a distance d21 between the contour point a21 and the internal point b21, a distance d22 between the contour point a21 and the internal point b22, a distance d23 between the contour point a21 and the internal point b23, and the contour point a21 and the internal point b24.
  • the distance d24 is calculated.
  • the distance calculation unit 506 calculates a distance that passes through the inside of the shape data sd21. This is because the distance represents the feature of the internal region of the shape data sd21.
  • FIG. 11 is a diagram illustrating an example of a list in which distances are rearranged.
  • the rearrangement unit 510 rearranges the distances calculated by the distance calculation unit 506 in ascending order, for example.
  • the list of distances rearranged by the rearrangement unit 510 is a geodesic distance profile.
  • FIG. 12 is a diagram illustrating an example of area division.
  • the area dividing unit 504 divides the shape data sd21 into a plurality of areas ar21, ar22, ar23, and ar24.
  • the area dividing unit 504 divides each area so that one internal point is included.
  • the profile generation unit 502 may set the internal points by dividing into a plurality of regions by the region dividing unit 504.
  • the area calculation unit 508 calculates the area of each divided area.
  • the area of the area ar21 is s1
  • the area of the area ar22 is s2
  • the area of the area ar23 is s3
  • the area of the area ar24 is s4.
  • FIG. 13 is a diagram showing an example of a list reflecting the area of each region.
  • rearrangement unit 510 adjusts the width of the horizontal axis for each internal point according to the area of the region having each internal point. That is, the width of the horizontal axis of each internal point is a length proportional to the size of the area having each internal point.
  • FIG. 14 is a diagram showing an example of the final result of the geodesic distance profile.
  • the profile pr21 shown in FIG. 14 is the final geodesic distance profile as a shape descriptor associated with the contour point a21.
  • FIG. 15 is a diagram illustrating an example of distance calculation in the case of using triangular mesh division.
  • the shape data sd31 is divided into triangular meshes by the area dividing unit 504.
  • the center of each triangle is set to each internal point b31, b32, b33, b34 by the area dividing unit 504.
  • the reason why the center of the triangle is the internal point is that the relationship between the internal point and the region can be easily handled in a one-to-one relationship.
  • the distance calculation unit 506 sets a path from the center of each triangle to each vertex of the triangle including the center and the center of another adjacent triangle. Note that the sides of each triangle are also set as paths.
  • the distance calculation unit 506 calculates the shortest path length from the contour point a31 to each internal point as the distance.
  • the distance calculation unit 506 sets the distance between the contour point a31 and the internal point b31 as the length of the path ps31, sets the distance between the contour point a31 and the internal point b32 as the length between the path ps31 and the path ps32, and sets the contour point a31. Is the length of the path ps33 and the path ps34, and the distance between the contour point a31 and the internal point b34 is the length of the path ps33 and the path ps35. Thereby, the distance calculation part 506 can calculate the distance which passes the inside of shape data comparatively easily.
  • FIG. 16 is a diagram for explaining an outline of matching processing in the embodiment using an image.
  • the input unit 202 inputs shape data sd31 and shape data sd41.
  • the setting unit 206 sets a geodesic distance profile for each sampled contour point.
  • the profile pr31 represents a geodesic distance profile of the contour point a31 of the shape data sd31.
  • the profile pr41 represents a geodesic distance profile of the contour point a41 of the shape data sd41.
  • the matching unit 208 calculates a logarithm with respect to the distance (d) of the geodetic distance profile.
  • the profile pr32 represents a geodetic distance profile after logarithm calculation
  • the profile pr42 represents a geodetic distance profile after logarithm calculation.
  • the matching unit 208 calculates a difference between geodetic distance profiles after logarithmic calculation, and performs matching processing of contour points in the two shape data based on this difference.
  • the difference includes, for example, a square error.
  • the matching unit 208 obtains contour points of other shape data in which the difference in geodesic distance profile is the smallest at a predetermined contour point of one shape data.
  • the contour point matching process considering the internal region of the shape data is performed. Further, by calculating the area of the region including the internal point and the logarithm of the distance (d), the information processing apparatus 10 can improve the matching accuracy.
  • FIG. 17 is a flowchart illustrating an example of processing from input of shape data to matching in the embodiment.
  • the input unit 202 inputs the first shape data and the second shape data.
  • the input unit 202 may input the first shape data serving as the source first and input the second shape data serving as the target later.
  • step S104 the sampling unit 204 samples the contour points in the first shape data and the contour points in the second shape data.
  • the sampling unit 204 samples the contour points so as to be as evenly spaced as possible.
  • step S106 the setting unit 206 sets the first shape descriptor at the contour point of the first shape data, and sets the second shape descriptor at the contour point of the second shape data.
  • the shape descriptor is, for example, a geodetic distance profile considering the internal region of the shape data as described above. Generation of the geodesic distance profile will be described with reference to FIG.
  • the matching unit 208 compares the first shape descriptor with the second shape descriptor, and associates each point in the contour of the first shape data with each point in the contour of the second shape data. I do. For example, the matching unit 208 obtains the contour point of the second shape data that minimizes the square error between the geodesic distance profiles having the logarithm calculated distances with respect to the contour point of the first shape data. The matching unit 208 obtains matching contour points of the second shape data for all contour points of the first shape data.
  • step S110 the output unit 210 outputs a set of the contour point of the first shape data and the contour point of the second shape data obtained by the matching process.
  • the information processing apparatus 10 can perform matching processing on the contour of the shape in consideration of the internal region of the shape.
  • FIG. 18 is a flowchart illustrating an example of geodetic distance profile generation processing in the embodiment.
  • the process shown in FIG. 18 shows a process of generating a geodesic distance profile at all contour points sampled with one shape data.
  • the region dividing unit 504 divides the shape data into a plurality of regions. For example, the region dividing unit 504 divides the shape data into a plurality of regions using triangular mesh division.
  • step S204 the area calculation unit 508 calculates the area of each divided region. For example, the area calculation unit 508 calculates the area of each triangle divided into triangle meshes.
  • step S206 the distance calculation unit 506 calculates the distance between any one contour point of the shape data and each internal point of the shape data.
  • the distance calculation unit 506 calculates a distance passing through the inside of the shape data.
  • step S208 the rearrangement unit 510 rearranges the distances at one contour point in ascending order to generate a list.
  • This list becomes a geodesic distance profile.
  • step S210 the rearrangement unit 510 reflects the area of each region in the geodesic distance profile.
  • the width of the axis of the internal point in the list is adjusted in proportion to the area of the region including the internal point. Thereby, a geodesic distance profile reflecting the area of each region is generated.
  • step S212 the profile generation unit 502 determines whether geodesic distance profiles have been generated at all contour points. If geodesic distance profiles have been generated at all contour points (step S212—YES), the profile generation process ends. If no geodetic distance profile has been generated at all contour points (step S212—NO), the process of step S206 is executed again to generate a geodetic distance profile at another contour point.
  • the profile generation unit 502 can generate a geodesic distance profile that suitably reflects the internal characteristics of the shape data.
  • FIG. 19 is a diagram illustrating an example of shape data used in Experiment 1.
  • FIG. 19A shows the first shape data as a source
  • FIG. 19B shows the second shape data as a target.
  • the shape data shown in FIG. 19 is shape data in a state where the triangular mesh is divided.
  • the information processing apparatus 10 determines which contour point of the second shape data matches the contour point of the first shape data.
  • FIG. 20 is a diagram showing the degree of distortion in the internal region when performing contour matching in Experiment 1.
  • FIG. 20A shows the state of the internal region of the first shape data.
  • FIG. 20B shows an internal region based on contour points matched according to the present embodiment. Referring to FIG. 20A and FIG. 20B, the heads and feet of the first shape data are appropriately associated with each other in the second shape data. Therefore, it is possible to prevent the distortion of the inner region.
  • the part indicated by the arrow shown in FIG. 20C is supposed to be associated with the outline of the eye part of the first shape data, but is actually associated with the outline of the neck part. This indicates that the matching process is not appropriate and the inner region is distorted when this matching process is used.
  • FIG. 21 is a diagram showing a texture that has been texture-transformed by contour matching in Experiment 1.
  • FIG. FIG. 21A shows a state in which an animal A (for example, a cat) is pasted on the first shape data.
  • FIG. 21B shows animal A texture-transformed according to this embodiment. Referring to FIGS. 21A and 21B, the outline of the first shape data is appropriately associated with the outline of the second shape data. Therefore, it is possible to prevent the distortion of the inner region.
  • the arrow part shown to FIG.21 (C) and FIG.21 (D) shows the location where matching of the contour point was not performed appropriately like FIG. In particular, in the example shown in FIG. 21D, most contour points are not appropriately matched.
  • FIG. 22 is a diagram illustrating an example of shape data used in Experiment 2.
  • FIG. 22A shows the first shape data as the source
  • FIG. 22B shows the second shape data as the target.
  • the shape data shown in FIG. 22 is shape data in a state where the triangular mesh is divided.
  • the information processing apparatus 10 determines which contour point of the second shape data matches the contour point of the first shape data.
  • FIG. 23 is a diagram showing the degree of distortion in the internal region when performing contour matching in Experiment 2.
  • FIG. 23A shows the state of the internal area of the first shape data.
  • FIG. 23B shows an internal region based on contour points matched according to the present embodiment. Referring to FIG. 23A and FIG. 23B, the heads and feet of the first shape data are appropriately associated also in the second shape data. Therefore, it is possible to prevent the distortion of the inner region.
  • the part indicated by the arrow in FIG. 23C is supposed to be associated with the outline of the toe part of the first shape data, but is actually associated with the outline of the tail part. This indicates that the matching process is not appropriate and the inner region is distorted when this matching process is used.
  • the part of the arrow shown in FIG. 23D is supposed to be associated with the outline of a part of the hind legs of the first shape data, but is actually associated with the outline of the front part of the hind legs. This indicates that the matching process is not appropriate and the inner region is distorted when this matching process is used.
  • FIG. 24 is a diagram showing a texture that has been texture-transformed by contour matching in Experiment 2.
  • FIG. FIG. 24A shows a state where the texture of animal B (for example, dog) is pasted on the first shape data.
  • FIG. 24B shows an animal B) that has been texture transformed according to this embodiment.
  • the contour of the first shape data is appropriately associated with the contour of the second shape data. Therefore, it is possible to prevent the distortion of the inner region.
  • the arrow part shown to FIG.24 (C) and FIG.24 (D) shows the location where matching of the contour point was not performed appropriately like FIG.
  • FIG. 25 is a diagram illustrating an example of shape data used in Experiment 3.
  • FIG. 25A shows the first shape data as the source
  • FIG. 25B shows the second shape data as the target.
  • a closed region exists inside the shape data.
  • FIG. 26 is a diagram showing the experimental results of Experiment 3.
  • FIG. 26A shows an internal region based on contour points matched by the conventional method. As shown in FIG. 25 (A) and FIG. 26 (A), since the arrows indicate different portions, it is understood that the contour matching of the closed region is not properly performed. This is because in the conventional method, the contour of the shape data and the contour of the internal closed region are matched independently.
  • FIG. 26B shows an internal region based on contour points matched by the method of the present embodiment. As shown in FIGS. 25 (A) and 26 (B), since the arrows indicate substantially the same part, it is understood that the contour matching of the closed region is appropriately performed. This is because, in the contour of the closed region, the feature of the internal region between the shape data and the closed region is used for matching as a shape descriptor.
  • FIG. 27 is a diagram illustrating experimental results of Experiment 4.
  • FIG. 27A shows shape data as a source and internal regions.
  • FIG. 27 (B) shows an internal region in the case of using target shape data and a geodesic distance profile without filtering. According to the arrow shown in FIG. 27 (B), the source shape data and the target shape data are matched with the left and right outlines.
  • FIG. 27 (C) shows an internal region in the case of using target shape data and a geodesic distance profile with low-pass filter processing.
  • contour matching is appropriately performed between the source shape data and the target shape data. This is because by passing a low-pass filter through the geodesic distance profile, the influence of the contour unevenness indicated by the Local portion becomes stronger in the matching process.
  • the contour matching process considering the internal region can be appropriately performed by using the geodesic distance profile as the shape descriptor considering the internal region.
  • the program executed by the information processing apparatus 10 As actual hardware, when the CPU 102 reads the program from the ROM 106 and executes it, one or more of the above-described units are loaded onto the RAM 104. One or more units are generated on the RAM 104.
  • the matching process described in the above-described embodiment may be realized as a program for causing a computer to execute.
  • the matching process described above can be realized by installing this program from a server or the like and causing the computer to execute it.
  • the recording medium 116 is a recording medium that records information optically, electrically, or magnetically, such as a CD-ROM, a flexible disk, or a magneto-optical disk, and information is electrically stored such as a ROM or flash memory.
  • Various types of recording media such as a semiconductor memory for recording can be used.
  • the description has been given by taking the two-dimensional shape data as an example.
  • a shape descriptor considering the internal region can be set as the contour point. Therefore, this embodiment can also be applied to contour matching between three-dimensional shape data.
  • the matching result processed by this embodiment is applied to the interpolation processing of shape data in the frame between the key frames when, for example, key frames of the first frame and the last frame are given in the animation. can do.
  • this matching result can be applied to deformation such as texture transformation that transfers the pattern of shape data, photo editing, or shape deformation. Further, this matching result can be applied as a search key or a classification identifier for shape classification when there are many shape data in the database.
  • This matching result can also be applied to animation coloring.
  • the shape data of the first frame can be painted, and the shape data of the second and subsequent frames can be automatically determined which color is painted on which contour portion using the matching result.
  • Information processing apparatus 102 CPU 104 RAM 106 ROM 202 Input unit 204 Sampling unit 206 Setting unit 208 Matching unit 210 Output unit 402 Enlargement / reduction unit 502 Profile generation unit 504 Area division unit 506 Distance calculation unit 508 Area calculation unit 510 Rearrangement unit 602 Logarithmic calculation unit 604 Filter processing unit

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Multimedia (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Evolutionary Computation (AREA)
  • Databases & Information Systems (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Processing Or Creating Images (AREA)
  • Image Analysis (AREA)

Abstract

An information processing device is provided with: an input unit to which first shape data and second shape data are inputted; a sampling unit which samples each of points in the contours of the first shape data and the second shape data; a setting unit which sets a first shape descriptor based on an inner region of the first shape data with respect to each of the points sampled in the first shape data, and a second shape descriptor based on an inner region of the second shape data with respect to each of the points sampled in the second shape data; a matching unit which compares the first shape descriptor and the second shape descriptor, and associates each of the points in the contour of the first shape data and each of the points in the contour of the second shape data; and an output unit which outputs sets of the associated points in the contour of the first shape data and points in the contour of the second shape data.

Description

情報処理装置、情報処理方法、プログラム及び記録媒体Information processing apparatus, information processing method, program, and recording medium
　本発明は、マッチング処理を行う情報処理装置、情報処理方法、プログラム及び記録媒体に関する。 The present invention relates to an information processing apparatus, an information processing method, a program, and a recording medium that perform matching processing.
　従来、異なる形状データの輪郭に対してマッチングを行う技術が数多く研究されてきた。一般的に用いられる２次元形状の輪郭マッチング方法では、輪郭線上での曲がり具合、例えば曲率などを用いてマッチングを行う技術がある（非特許文献１）。 Conventionally, many techniques for matching contours of different shape data have been studied. As a generally used two-dimensional contour matching method, there is a technique for performing matching using a degree of bending on a contour line, for example, a curvature (Non-Patent Document 1).
　また、３次元形状においては、例えば、３次元形状の２点間の表面上の距離を用いてマッチングを行う技術や、３次元形状表面の２点間のユークリッド距離を用いてマッチングを行う技術がある（非特許文献２、３）。 In the 3D shape, for example, there is a technology that performs matching using the distance on the surface between two points of the 3D shape and a technology that performs matching using the Euclidean distance between two points on the surface of the 3D shape. (Non-Patent Documents 2 and 3).
　しかしながら、従来技術では、輪郭の曲がり具合や表面形状などを用いてマッチングを行うため、マッチング後の輪郭を用いてテクスチャトランスフォームなどを行うと、内部領域が歪んでいることが多い。これは、内部領域を考慮せずに、輪郭線や表面形状の点を用いてマッチングを行うことに起因すると考えられる。 However, in the prior art, since matching is performed using the curved shape of the contour, the surface shape, and the like, the inner region is often distorted when texture transformation is performed using the contour after matching. This is considered to be caused by performing matching using a contour line or a surface shape point without considering the internal region.
　そこで、本発明の所定の態様は、上記課題に鑑みてなされたものであり、形状の内部領域を考慮した、形状の輪郭に対するマッチング処理を行うことができる情報処理装置、情報処理方法、プログラム及び記録媒体を提供することを目的とする。 Therefore, a predetermined aspect of the present invention has been made in view of the above problems, and an information processing apparatus, an information processing method, a program, and a program capable of performing matching processing on a contour of a shape in consideration of an inner region of the shape An object is to provide a recording medium.
　本発明の一態様における情報処理装置は、第１形状データと第２形状データとを入力する入力部と、前記第１形状データ及び前記第２形状データの輪郭における各点をサンプリングするサンプリング部と、前記第１形状データにおいてサンプリングされた各点に対し、前記第１形状データの内部領域に基づく第１形状記述子と、前記第２形状データにおいてサンプリングされた各点に対し、前記第２形状データの内部領域に基づく第２形状記述子とを設定する設定部と、前記第１形状記述子と前記第２形状記述子とを比較して、前記第１形状データの輪郭における各点と、前記第２形状データの輪郭における各点とを対応付けるマッチング部と、対応付けられた前記第１形状データの輪郭における点と、前記第２形状データの輪郭における点とのセットを出力する出力部と、を備える。これにより、形状の内部領域を考慮した、形状の輪郭に対するマッチング処理を行うことができる。 An information processing apparatus according to an aspect of the present invention includes an input unit that inputs first shape data and second shape data, a sampling unit that samples each point in the outline of the first shape data and the second shape data, and For each point sampled in the first shape data, a first shape descriptor based on an internal region of the first shape data, and for each point sampled in the second shape data, the second shape A setting unit for setting a second shape descriptor based on an internal area of the data, comparing the first shape descriptor and the second shape descriptor, and each point in the contour of the first shape data; A matching unit that associates each point in the contour of the second shape data, a point in the contour of the first shape data that is associated, and a contour in the contour of the second shape data And an output unit for outputting a set of. Thereby, the matching process with respect to the outline of a shape in consideration of the internal region of the shape can be performed.
　また、前記設定部は、前記第１形状記述子として、前記サンプリングされた点と前記第１形状データ内部の各点との、前記第１形状データ内部を通る距離を並べ替えた第１測地的距離プロファイルを生成し、前記第２形状記述子として、前記サンプリングされた点と前記第２形状データ内部の各点との、前記第２形状データ内部を通る距離を並べ替えた第２測地的距離プロファイルを生成してもよい。これにより、形状記述子として適切な測地的距離プロファイルを生成することができる。 In addition, the setting unit, as the first shape descriptor, a first geodetic that rearranges the distance through the first shape data between the sampled point and each point in the first shape data A second geodesic distance obtained by generating a distance profile and rearranging the distance passing through the inside of the second shape data between the sampled point and each point inside the second shape data as the second shape descriptor A profile may be generated. Thereby, a geodesic distance profile suitable as a shape descriptor can be generated.
　また、前記設定部は、前記第１形状データ及び前記第２形状データを複数の領域に分割し、各領域内の点を内部の点としてもよい。これにより、適切な内部点の設定を容易にすることができる。 Further, the setting unit may divide the first shape data and the second shape data into a plurality of regions, and use points in each region as internal points. Thereby, the setting of a suitable internal point can be made easy.
　また、前記設定部は、三角形メッシュ分割を用いて、前記第１形状データ及び前記第２形状データを複数の領域に分割し、前記各領域内の点を、分割された三角形の中心点としてもよい。これにより、領域と内部点との対応関係を明確にすることができる。 Further, the setting unit divides the first shape data and the second shape data into a plurality of regions by using a triangular mesh division, and a point in each region may be used as a center point of the divided triangle. Good. Thereby, the correspondence between the region and the internal point can be clarified.
　また、前記設定部は、前記各領域の面積を考慮して前記距離を並べ替えることにより、前記第１測地的距離プロファイル及び前記第２測地的距離プロファイルを生成してもよい。これにより、内部領域の特徴をより反映した測地的距離プロファイルを生成することができる。 Further, the setting unit may generate the first geodesic distance profile and the second geodesic distance profile by rearranging the distances in consideration of the area of each region. Thereby, it is possible to generate a geodesic distance profile that more reflects the characteristics of the internal region.
　また、前記マッチング部は、前記第１測地的距離プロファイルの対数と前記第２測地的距離プロファイルの対数とを計算し、計算された対数同士の差分の合計が最も小さくなるように対応付け処理を行ってもよい。これにより、マッチングの精度をさらに向上させることができる。 The matching unit calculates a logarithm of the first geodetic distance profile and a logarithm of the second geodetic distance profile, and performs an association process so that a sum of differences between the calculated logarithms is minimized. You may go. Thereby, the matching accuracy can be further improved.
　また、前記マッチング部は、前記第１測地的距離プロファイルと、前記第２測地的距離プロファイルとにフィルタ処理を行い、フィルタ処理後の各測地的距離プロファイルの差分が最も小さくなるように対応付け処理を行ってもよい。これにより、用途に応じた輪郭のマッチング処理を行うことができる。 The matching unit performs a filtering process on the first geodetic distance profile and the second geodetic distance profile, and performs an association process so that a difference between the geodetic distance profiles after the filtering process is minimized. May be performed. Thereby, the outline matching process according to a use can be performed.
　また、前記第１形状データ内に第１閉領域があり、前記第２形状データ内に第２閉領域がある場合、前記設定部は、前記第１形状データと前記第１閉領域との間にある領域に基づいて、前記第１閉領域の輪郭における各点の第３形状記述子を設定し、前記第２形状データと前記第２閉領域との間にある領域に基づいて、前記第２閉領域の輪郭における各点の第４形状記述子を設定してもよい。これにより、形状データの内部に閉領域がある場合に、この閉領域の輪郭についても適切なマッチング処理を行うことができる。 In addition, when the first shape data includes a first closed region and the second shape data includes the second closed region, the setting unit may determine whether the first shape data is between the first shape data and the first closed region. A third shape descriptor of each point in the contour of the first closed region is set based on the region in the first closed region, and the second shape data and the second closed region are used to determine the third shape descriptor. You may set the 4th shape descriptor of each point in the outline of 2 closed areas. Thereby, when there is a closed region in the shape data, an appropriate matching process can be performed for the contour of the closed region.
　また、前記サンプリング部は、前記第１形状データの内部面積と前記第２形状データの内部面積とが一致するように、前記第１形状データ又は前記第２形状データを拡大又は縮小してもよい。これにより、マッチング対象の形状データのサイズが異なる場合であっても、適切に輪郭マッチング処理を行うことができる。 The sampling unit may enlarge or reduce the first shape data or the second shape data so that the internal area of the first shape data matches the internal area of the second shape data. . Thereby, even if the size of the shape data to be matched is different, the contour matching process can be appropriately performed.
　また、本発明の他の態様における情報処理方法は、コンピュータが実行する情報処理方法であって、第１形状データと第２形状データとを入力するステップと、前記第１形状データ及び前記第２形状データの輪郭における各点をサンプリングするステップと、前記第１形状データにおいてサンプリングされた各点に対し、前記第１形状データの内部領域に基づく第１形状記述子と、前記第２形状データにおいてサンプリングされた各点に対し、前記第２形状データの内部領域に基づく第２形状記述子とを設定するステップと、前記第１形状記述子と前記第２形状記述子とを比較して、前記第１形状データの輪郭における各点と、前記第２形状データの輪郭における各点とを対応付けるステップと、対応付けられた前記第１形状データの輪郭における点と、前記第２形状データの輪郭における点とのセットを出力するステップと、を含む。これにより、上記の情報処理装置と同様に、形状の内部領域を考慮した、形状の輪郭に対するマッチング処理を行うことができる。 An information processing method according to another aspect of the present invention is an information processing method executed by a computer, the step of inputting first shape data and second shape data, the first shape data, and the second shape data. Sampling each point in the contour of the shape data; for each point sampled in the first shape data, a first shape descriptor based on an internal region of the first shape data; and in the second shape data For each sampled point, setting a second shape descriptor based on an internal region of the second shape data; comparing the first shape descriptor and the second shape descriptor; and Associating each point in the contour of the first shape data with each point in the contour of the second shape data, and the contour of the associated first shape data Including a point where definitive, and outputting a set of points in the second shape data profile, a. Thereby, the matching process with respect to the outline of a shape which considered the internal area | region of a shape can be performed similarly to said information processing apparatus.
　本発明の所定の態様によれば、形状の内部領域を考慮した、形状の輪郭に対するマッチング処理を行うことができる。 According to the predetermined aspect of the present invention, it is possible to perform the matching process on the contour of the shape in consideration of the internal region of the shape.
実施形態における情報処理装置の概略構成の一例を示すブロック図である。It is a block diagram which shows an example of schematic structure of the information processing apparatus in embodiment. 実施形態における情報処理装置の機能の一例を示すブロック図である。It is a block diagram which shows an example of the function of the information processing apparatus in embodiment. 実施形態におけるＲＡＭに含まれる各記憶部の例を示す図である。It is a figure which shows the example of each memory | storage part contained in RAM in embodiment. 実子形態におけるサンプリング部の構成の一例を示すブロック図である。It is a block diagram which shows an example of a structure of the sampling part in a real child form. 実施形態における設定部の構成の一例を示すブロック図である。It is a block diagram which shows an example of a structure of the setting part in embodiment. 実施形態における測地的距離プロファイルを説明するための図である。It is a figure for demonstrating the geodesic distance profile in embodiment. 実施形態におけるマッチング部の構成の一例を示すブロック図である。It is a block diagram which shows an example of a structure of the matching part in embodiment. 形状データの一例を示す図である。It is a figure which shows an example of shape data. 複数の内部点の一例を示す図である。It is a figure which shows an example of several internal points. 輪郭点と内部点との距離の一例を示す図である。It is a figure which shows an example of the distance of an outline point and an internal point. 距離を並べ替えたリストの一例を示す図である。It is a figure which shows an example of the list | wrist which rearranged distance. 領域分割の一例を示す図である。It is a figure which shows an example of area division. 各領域の面積を反映したリストの一例を示す図である。It is a figure which shows an example of the list reflecting the area of each area | region. 測地的距離プロファイルの最終結果の一例を示す図である。It is a figure which shows an example of the final result of geodesic distance profile. 三角形メッシュ分割を用いた場合の距離の計算の一例を示す図である。It is a figure which shows an example of calculation of the distance at the time of using a triangular mesh division | segmentation. 実施形態におけるマッチング処理までの概要を、イメージを用いて説明するための図である。It is a figure for demonstrating the outline | summary to the matching process in embodiment using an image. 実施形態における形状データの入力からマッチングまでの処理の一例を示すフローチャートである。It is a flowchart which shows an example of the process from the input of the shape data in embodiment to a matching. 実施形態における測地的距離プロファイルの生成処理の一例を示すフローチャートである。It is a flowchart which shows an example of the production | generation process of the geodesic distance profile in embodiment. 実験１に用いた形状データの一例を示す図である。It is a figure which shows an example of the shape data used for Experiment. 実験１における輪郭マッチングを行う際の内部領域の歪み具合を示す図である。It is a figure which shows the distortion condition of the internal area | region at the time of performing the outline matching in Experiment 1. FIG. 実験１における輪郭マッチングによりテクスチャトランスフォームされたテクスチャを示す図である。It is a figure which shows the texture texture-transformed by the outline matching in Experiment 1. FIG. 実験２に用いた形状データの一例を示す図である。It is a figure which shows an example of the shape data used for Experiment. 実験２における輪郭マッチングを行う際の内部領域の歪み具合を示す図である。It is a figure which shows the distortion condition of the internal area | region at the time of performing the outline matching in Experiment 2. FIG. 実験２における輪郭マッチングによりテクスチャトランスフォームされたテクスチャを示す図である。It is a figure which shows the texture texture-transformed by the outline matching in Experiment 2. FIG. 実験３に用いる形状データの一例を示す図である。It is a figure which shows an example of the shape data used for Experiment 3. 実験３の実験結果を示す図である。It is a figure which shows the experimental result of the experiment 3. FIG. 実験４の実験結果を示す図である。It is a figure which shows the experimental result of the experiment 4. FIG.
　以下、添付図面を参照しながら本発明の実施形態について詳細に説明する。 Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
　［実施形態］
　＜構成＞
　図１は、実施形態における情報処理装置１０の概略構成の一例を示すブロック図である。図１に示すように、情報処理装置１０は、ＣＰＵ（Central Processing Unit）１０２と、ＲＡＭ（Random Access Memory）１０４と、ＲＯＭ（Read only Memory）１０６と、ドライブ装置１０８と、ネットワークＩ／Ｆ（Interface）１１０と、入力装置１１２と、及び表示装置１１４とを有する。これら各構成は、バスを介して相互にデータ送受信可能に接続されている。 [Embodiment]
<Configuration>
FIG. 1 is a block diagram illustrating an example of a schematic configuration of an information processing apparatus 10 according to the embodiment. As shown in FIG. 1, the information processing apparatus 10 includes a central processing unit (CPU) 102, a random access memory (RAM) 104, a read only memory (ROM) 106, a drive device 108, a network I / F ( Interface) 110, an input device 112, and a display device 114. These components are connected to each other via a bus so as to be able to transmit and receive data.
　ＣＰＵ１０２は、コンピュータの中で、各装置の制御やデータの演算、加工を行う制御部である。また、ＣＰＵ１０２は、ＲＡＭ１０４又はＲＯＭ１０６に記憶されたマッチング処理のプログラムを実行する演算装置である。ＣＰＵ１０２は、入力装置１１２やネットワークＩ／Ｆ１１０などから形状データを受け取り、演算、加工した上で、演算結果を表示装置１１４や記憶装置などに出力する。 The CPU 102 is a control unit that controls each device, calculates data, and processes in the computer. The CPU 102 is an arithmetic unit that executes a matching processing program stored in the RAM 104 or the ROM 106. The CPU 102 receives shape data from the input device 112, the network I / F 110, and the like, calculates and processes the data, and outputs the calculation result to the display device 114, the storage device, and the like.
　ＲＡＭ１０４は、例えば主記憶部などである。ＲＡＭ１０４は、ＣＰＵ１０２が実行する基本ソフトウェアであるＯＳ（Operating System）やアプリケーションソフトウェアなどのプログラムやデータを記憶又は一時保存する記憶装置である。 The RAM 104 is, for example, a main storage unit. The RAM 104 is a storage device that stores or temporarily stores programs and data such as OS (Operating System) and application software that are basic software executed by the CPU 102.
　ＲＯＭ１０６は、例えばアプリケーションソフトウェアなどに関連するデータを記憶する記憶装置である。 The ROM 106 is a storage device that stores data related to application software, for example.
　ドライブ装置１０８は、記録媒体１１６、例えばＣＤ－ＲＯＭやＳＤカードなどからプログラムを読み出し、記憶装置にインストールする。 The drive device 108 reads the program from the recording medium 116, such as a CD-ROM or SD card, and installs it in the storage device.
　また、記録媒体１１６に、所定のプログラムを格納し、この記録媒体１１６に格納されたプログラムはドライブ装置１０８を介して情報処理装置１０にインストールされる。インストールされた所定のプログラムは、情報処理装置１０により実行可能となる。 Further, a predetermined program is stored in the recording medium 116, and the program stored in the recording medium 116 is installed in the information processing apparatus 10 via the drive device 108. The installed predetermined program can be executed by the information processing apparatus 10.
　ネットワークＩ／Ｆ１１０は、通信機能を有する周辺機器と情報処理装置１０とのインターフェースである。また、ネットワークＩ／Ｆ１１０は、例えば、有線及び／又は無線回線などのデータ伝送路により構築されたＬＡＮ（Local Area Network）、ＷＡＮ（Wide Area Network）などのネットワークを介して接続される。 The network I / F 110 is an interface between a peripheral device having a communication function and the information processing apparatus 10. The network I / F 110 is connected via a network such as a LAN (Local Area Network) or a WAN (Wide Area Network) constructed by a data transmission path such as a wired and / or wireless line.
　入力装置１１２は、カーソルキー、数字入力及び各種機能キー等を備えたキーボード、表示装置１１４の表示画面上でキーの選択等を行うためのマウスやスライドパッド等を有する。また、入力装置１１２は、ユーザがＣＰＵ１０２に操作指示を与えたり、データを入力したりするためのユーザインターフェースである。 The input device 112 includes a keyboard having cursor keys, numeric input, various function keys, and the like, a mouse and a slide pad for selecting keys on the display screen of the display device 114, and the like. The input device 112 is a user interface for a user to give an operation instruction to the CPU 102 or input data.
　表示装置１１４は、ＬＣＤ（Liquid Crystal Display）等により構成され、ＣＰＵ１０２から入力される表示データに応じた表示が行われる。なお、入力装置１１２や表示装置１１４は、情報処理装置１０の外部に設けられてもよい。 The display device 114 is configured by an LCD (Liquid Crystal Display) or the like, and performs display according to display data input from the CPU 102. Note that the input device 112 and the display device 114 may be provided outside the information processing apparatus 10.
　＜機能＞
　次に、情報処理装置１０の機能について説明する。図２は、実施形態における情報処理装置１０の機能の一例を示すブロック図である。図２に示す情報処理装置１０は、入力部２０２と、サンプリング部２０４と、設定部２０６と、マッチング部２０８と、出力部２１０とを少なくとも有する。 <Function>
Next, functions of the information processing apparatus 10 will be described. FIG. 2 is a block diagram illustrating an example of functions of the information processing apparatus 10 according to the embodiment. The information processing apparatus 10 illustrated in FIG. 2 includes at least an input unit 202, a sampling unit 204, a setting unit 206, a matching unit 208, and an output unit 210.
　なお、図２に示す各部は、例えばＣＰＵ１０２やワークメモリとしてのＲＡＭ１０４などにより実現されうる。ＣＰＵ１０２は、マッチング処理のプログラムを実行することで、各部の機能を実行することができる。 2 can be realized by the CPU 102 or the RAM 104 as a work memory, for example. The CPU 102 can execute the function of each unit by executing a matching processing program.
　入力部２０２は、ＲＡＭ１０４、ＲＯＭ１０６、ネットワークＩ／Ｆ１１０、又は入力装置１１２から、ソース（source）の形状データとターゲット（target）の形状データとを入力する。以下では、ソースの形状データを第１形状データとも称し、ターゲットの形状データを第２形状データとも称する。入力部２０２は、入力された第１形状データ及び第２形状データをサンプリング部２０４に出力する。入力部２０２は、ソースとなる第１形状データを先に入力し、ターゲットとなる第２形状データを後に入力するようにしてもよい。 The input unit 202 inputs source shape data and target shape data from the RAM 104, the ROM 106, the network I / F 110, or the input device 112. Hereinafter, the source shape data is also referred to as first shape data, and the target shape data is also referred to as second shape data. The input unit 202 outputs the input first shape data and second shape data to the sampling unit 204. The input unit 202 may input the first shape data serving as the source first and input the second shape data serving as the target later.
　サンプリング部２０４は、第１形状データ及び第２形状データの輪郭における各点をサンプリングする。サンプリング部２０４は、入力部２０２から第１形状データ及び第２形状データを取得すると、それぞれの形状データの輪郭において、複数の点をサンプリングする。 The sampling unit 204 samples each point in the outline of the first shape data and the second shape data. When acquiring the first shape data and the second shape data from the input unit 202, the sampling unit 204 samples a plurality of points in the contour of each shape data.
　サンプリング部２０４は、例えば、第１形状データ及び第２形状データの輪郭において、等間隔となるように輪郭点をサンプリングしていく。輪郭点とは、輪郭における点をいう。サンプリング部２０４は、第１形状データにおいてサンプリングした輪郭点と、第２形状データにおいてサンプリングされた輪郭点とをマッチング部２０８に出力する。 The sampling unit 204 samples the contour points so as to be equally spaced in the contours of the first shape data and the second shape data, for example. A contour point refers to a point on the contour. The sampling unit 204 outputs the contour points sampled in the first shape data and the contour points sampled in the second shape data to the matching unit 208.
　設定部２０６は、第１形状データにおいてサンプリングされた各点に対し、第１形状データの内部領域に基づく第１形状記述子と、第２形状データにおいてサンプリングされた各点に対し、第２形状データの内部領域に基づく第２形状記述子とを設定する。内部領域に基づく形状記述子とは、例えば、輪郭点と、形状データ内にある複数の各点との距離の集合をいう。 For each point sampled in the first shape data, the setting unit 206 uses the first shape descriptor based on the internal region of the first shape data and the second shape for each point sampled in the second shape data. A second shape descriptor based on the internal area of the data is set. The shape descriptor based on the internal region refers to, for example, a set of distances between a contour point and a plurality of points in the shape data.
　設定部２０６は、例えば、形状データの内部領域に複数の点（以下、内部点ともいう）を設け、サンプリングされた輪郭点と各内部点との各距離を並べ替えた測地的距離プロファイル（geodesic distance profile）を、形状記述子（shape descriptor）とする。このとき、設定部２０６は、輪郭点と内部点との距離を求める際、直線距離などではなく、形状の内部を通る距離を求める。 The setting unit 206, for example, provides a plurality of points (hereinafter also referred to as internal points) in the internal region of the shape data, and rearranges the distances between the sampled contour points and the internal points (geodesic). Let distance profile be a shape descriptor. At this time, when determining the distance between the contour point and the internal point, the setting unit 206 determines the distance passing through the inside of the shape, not the linear distance.
　設定部２０６は、第１形状データに対する第１測地的距離プロファイルを第１形状記述とし、第２形状データに対する第２測地的距離プロファイルを第２形状記述子とする。設定部２０６は、設定した第１形状記述子と、第２形状記述子とをマッチング部２０８に出力する。 The setting unit 206 uses the first geodesic distance profile for the first shape data as the first shape description and the second geodesic distance profile for the second shape data as the second shape descriptor. The setting unit 206 outputs the set first shape descriptor and second shape descriptor to the matching unit 208.
　マッチング部２０８は、第１形状記述子と第２形状記述子とを比較して、第１形状データの輪郭における各点と、第２形状データの輪郭における各点とを対応付ける処理を行う。本実施形態において、マッチング処理とは、このような対応付け処理を含む。 The matching unit 208 compares the first shape descriptor and the second shape descriptor, and performs processing for associating each point in the contour of the first shape data with each point in the contour of the second shape data. In the present embodiment, the matching process includes such an association process.
　マッチング部２０８は、例えば、設定部２０６から取得した第１形状記述子と、第２形状記述子との差分が最も小さくなるように、第１形状データの輪郭点と第２形状データの輪郭点とのマッチング処理を行う。 The matching unit 208, for example, the contour point of the first shape data and the contour point of the second shape data so that the difference between the first shape descriptor acquired from the setting unit 206 and the second shape descriptor is minimized. Matching process is performed.
　効率よくマッチング処理を行うため、マッチング部２０８は、動的計画法を用いて、第１形状データの輪郭点とマッチングする第２形状データの輪郭点を求めていけばよい。マッチング部２０８は、マッチングされた第１形状データの輪郭点と第２形状データの輪郭点とを組（セット）にし、出力部２１０に出力する。輪郭点は、例えば座標データなどである。 In order to perform the matching process efficiently, the matching unit 208 may obtain the contour point of the second shape data that matches the contour point of the first shape data by using dynamic programming. The matching unit 208 sets the matched contour point of the first shape data and the contour point of the second shape data, and outputs the set to the output unit 210. The contour point is, for example, coordinate data.
　出力部２１０は、マッチングされた第１形状データの輪郭点と、第２形状データの輪郭点とのセットを出力する。出力部２１０は、例えば、第１形状データの全ての輪郭点において、それぞれマッチングされた第２形状データの輪郭点をセットにし、後段の処理部やＲＡＭ２０４又はＲＯＭ２０６に出力する。 The output unit 210 outputs a set of matched contour points of the first shape data and contour points of the second shape data. The output unit 210 sets, for example, the contour points of the second shape data matched with each of the contour points of the first shape data, and outputs the set to the subsequent processing unit, the RAM 204, or the ROM 206.
　次に、ＲＡＭ１０４に記憶されるデータについて説明する。図３は、実施形態におけるＲＡＭ１０４に含まれる各記憶部の例を示す図である。図３に示すＲＡＭ１０４は、形状データ記憶部３０２と、プロファイル記憶部３０４と、結果記憶部３０６とを含む。 Next, data stored in the RAM 104 will be described. FIG. 3 is a diagram illustrating an example of each storage unit included in the RAM 104 according to the embodiment. 3 includes a shape data storage unit 302, a profile storage unit 304, and a result storage unit 306.
　形状データ記憶部３０２は、入力部２０２により入力された第１形状データと、第２形状データとを記憶する。形状データ記憶部３０２は、第１形状データを有する第１画像と、第２形状データとを有する第２画像とを記憶してもよい。 The shape data storage unit 302 stores the first shape data and the second shape data input by the input unit 202. The shape data storage unit 302 may store a first image having first shape data and a second image having second shape data.
　プロファイル記憶部３０４は、第１形状データの各輪郭点に対する第１測地的距離プロファイルと、第２形状データの各輪郭点に対する第２測地的距離プロファイルとを記憶する。測地的距離プロファイルの詳細は後述する。 The profile storage unit 304 stores a first geodesic distance profile for each contour point of the first shape data and a second geodesic distance profile for each contour point of the second shape data. Details of the geodesic distance profile will be described later.
　結果記憶部３０６は、マッチング処理の結果を記憶する。マッチング処理の結果とは、第１形状データの各輪郭点に対し、マッチングされた第２形状データの輪郭点をセットにしたものである。 The result storage unit 306 stores the result of the matching process. The result of the matching process is a set of the contour points of the second shape data that have been matched for each contour point of the first shape data.
　以上、上述した処理を行う情報処理装置１０は、各輪郭点に測地的距離プロファイルを対応付けて、この測地的距離プロファイルを用いて輪郭点のマッチング処理を行うことで、内部領域の歪みを防止するマッチング処理を行うことができる。 As described above, the information processing apparatus 10 that performs the processing described above associates a geodesic distance profile with each contour point, and performs contour point matching processing using the geodesic distance profile, thereby preventing distortion of the internal region. Matching processing can be performed.
　＜各部の処理例＞
　次に、サンプリング部２０４、設定部２０６、及びマッチング部２０８の処理の一例について説明する。 <Processing example of each part>
Next, an example of processing of the sampling unit 204, the setting unit 206, and the matching unit 208 will be described.
　図４は、実施形態におけるサンプリング部２０４の構成の一例を示すブロック図である。図４に示すサンプリング部２０４は、拡大縮小部４０２を含む。拡大縮小部４０２は、第１形状データの内部面積と第２形状データの内部面積とが一致するように、第１形状データ又は第２形状データを拡大又は縮小する。 FIG. 4 is a block diagram illustrating an example of the configuration of the sampling unit 204 in the embodiment. The sampling unit 204 illustrated in FIG. 4 includes an enlargement / reduction unit 402. The enlargement / reduction unit 402 enlarges or reduces the first shape data or the second shape data so that the internal area of the first shape data matches the internal area of the second shape data.
　拡大縮小部４０２は、例えば、第１形状データ及び第２形状データの内部面積をそれぞれ計算し、第１形状データの面積と一致するように、第２形状データの面積を拡大又は縮小する。サンプリング部２０４は、拡大又は縮小された第２形状データに対して輪郭点のサンプリングを行う。 The enlargement / reduction unit 402 calculates, for example, the internal areas of the first shape data and the second shape data, respectively, and enlarges or reduces the area of the second shape data so as to match the area of the first shape data. The sampling unit 204 samples contour points for the enlarged or reduced second shape data.
　なお、拡大縮小部４０２は、サンプリング部２０４の一部として説明したが、サンプリング部２０４とは別の機能として入力部２０２と、サンプリング部２０４との間に設けてもよい。拡大縮小部４０２を設けることにより、情報処理装置１０は、マッチング対象の形状データのサイズが異なる場合であっても、適切に輪郭マッチング処理を行うことができる。また、拡大縮小部４０２で行った処理の逆処理は、後段の処理において必要であれば、適切なタイミングで実行されればよい。 The enlargement / reduction unit 402 has been described as a part of the sampling unit 204, but may be provided between the input unit 202 and the sampling unit 204 as a function different from the sampling unit 204. By providing the enlargement / reduction unit 402, the information processing apparatus 10 can appropriately perform contour matching processing even when the size of the shape data to be matched is different. Further, the reverse process of the process performed by the enlargement / reduction unit 402 may be executed at an appropriate timing if necessary in the subsequent process.
　図５は、実施形態における設定部２０６の構成の一例を示すブロック図である。図５に示すように、設定部２０６は、プロファイル生成部５０２を含み、プロファイル生成部５０２は、領域分割部５０４と、距離計算部５０６と、面積計算部５０８と、並べ替え部５１０とを含む。
　プロファイル生成部５０２は、以下の処理を行うことで、形状記述子として、サンプリングされた輪郭点と形状データの各内部点との、形状内部を通る距離を並べ替えた測地的距離プロファイルを生成する。 FIG. 5 is a block diagram illustrating an example of the configuration of the setting unit 206 in the embodiment. As illustrated in FIG. 5, the setting unit 206 includes a profile generation unit 502, and the profile generation unit 502 includes an area division unit 504, a distance calculation unit 506, an area calculation unit 508, and a rearrangement unit 510. .
The profile generation unit 502 generates a geodesic distance profile by rearranging the distances through the inside of the shape between the sampled contour point and each internal point of the shape data as a shape descriptor by performing the following processing. .
　領域分割部５０４は、取得した形状データを複数の領域に分割する。例えば、領域分割部５０４は、三角形メッシュ分割を用いて形状データを三角形メッシュに分割する。領域分割部５０４は、三角形メッシュ分割を用いることで、後述する領域の面積や距離が計算しやすくなる。領域分割部５０４は、分割した領域の情報を距離計算部５０６と、面積計算部５０８とに出力する。 The area dividing unit 504 divides the acquired shape data into a plurality of areas. For example, the region dividing unit 504 divides the shape data into triangular meshes using triangular mesh division. The area dividing unit 504 makes it easy to calculate the area and distance of an area described later by using triangular mesh division. The area dividing unit 504 outputs information on the divided areas to the distance calculating unit 506 and the area calculating unit 508.
　距離計算部５０６は、第１形状データの輪郭点ごとに、各内部点との距離を計算する。距離計算部５０６は、距離を求める際、内部領域を考慮するため、形状の内部を通る距離を計算する。距離計算部５０６は、第２形状データの輪郭点においても、各内部点との距離を計算する。三角形メッシュ分割を用いる場合の距離の計算方法については図１５を用いて後述する。距離計算部５０６は、第１形状データ及び第２形状データにおいて計算した、輪郭点ごとの各内部点との各距離を並べ替え部５１０に出力する。 The distance calculation unit 506 calculates the distance to each internal point for each contour point of the first shape data. The distance calculation unit 506 calculates the distance passing through the inside of the shape in order to consider the internal region when determining the distance. The distance calculation unit 506 calculates the distance to each internal point even in the contour point of the second shape data. A method of calculating the distance when using the triangular mesh division will be described later with reference to FIG. The distance calculation unit 506 outputs each distance to each internal point for each contour point calculated in the first shape data and the second shape data to the rearrangement unit 510.
　面積計算部５０８は、取得した形状データの各領域の面積を計算する。例えば、面積計算部５０８は、形状データが三角形メッシュ分割されている場合、形状データ内の各三角形の面積を計算する。面積計算部５０８は、第１形状データ内の各領域の面積と、第２形状データ内の各領域の面積とを並べ替え部５１０に出力する。 The area calculation unit 508 calculates the area of each area of the acquired shape data. For example, when the shape data is divided into triangle meshes, the area calculation unit 508 calculates the area of each triangle in the shape data. The area calculation unit 508 outputs the area of each region in the first shape data and the area of each region in the second shape data to the rearrangement unit 510.
　並べ替え部５１０は、各輪郭点において計算された各距離を所定の基準に従って並べ替える。例えば、並べ替え部５１０は、小さい（短い）順に各距離を並べ替えて、縦軸を距離、横軸を内部点数としてリストを生成する。このリストが、測地的距離プロファイルとなる。 The rearrangement unit 510 rearranges the distances calculated at the contour points according to a predetermined standard. For example, the rearrangement unit 510 rearranges the distances in ascending (short) order, and generates a list with the vertical axis as the distance and the horizontal axis as the internal score. This list becomes a geodesic distance profile.
　また、並べ替え部５１０は、このリストに対し、内部点の面積に比例して、横軸の幅を調整する。横軸の幅が調整された場合、横軸は、形状データの内部面積を表す。各領域の面積に応じてリストが生成されることで、プロファイル生成部５０２は、内部領域をより考慮した測地的距離プロファイルを生成することができる。 Also, the rearrangement unit 510 adjusts the width of the horizontal axis in proportion to the area of the internal point for this list. When the width of the horizontal axis is adjusted, the horizontal axis represents the internal area of the shape data. By generating a list according to the area of each region, the profile generation unit 502 can generate a geodesic distance profile that further considers the internal region.
　ここで、プロファイル生成部５０２により生成される測地的距離プロファイルについて図６を用いて説明する。図６は、実施形態における測地的距離プロファイルを説明するための図である。図６（Ａ）は、形状データ、及び輪郭点の一例を示す図である。図６（Ａ）に示す点ａ１１、ａ１２、ａ１３、ａ１４は、形状データｓｄ１１の輪郭において、サンプリング部２０４によりサンプリングされた複数の輪郭点を表す。図６（Ａ）に示す形状データｓｄ１１の各輪郭点に、測地的距離プロファイルが設定される。図６（Ａ）には、測地的距離プロファイルを設定する際に設けられる複数の内部点は図示していない。 Here, the geodetic distance profile generated by the profile generation unit 502 will be described with reference to FIG. FIG. 6 is a diagram for explaining a geodesic distance profile in the embodiment. FIG. 6A is a diagram illustrating an example of shape data and contour points. Points a11, a12, a13, and a14 shown in FIG. 6A represent a plurality of contour points sampled by the sampling unit 204 in the contour of the shape data sd11. A geodesic distance profile is set at each contour point of the shape data sd11 shown in FIG. FIG. 6A does not show a plurality of internal points provided when setting the geodesic distance profile.
　例えば、輪郭点ａ１１における測地的距離プロファイルは、輪郭点ａ１１から各内部点までの形状データｓｄ１１内部を通る各距離を並べ替えた集合である。上述したように、測地的距離プロファイルは、形状データｓｄ１１内部の各領域の面積を考慮して生成されるのが好適である。 For example, the geodesic distance profile at the contour point a11 is a set in which the distances passing through the shape data sd11 from the contour point a11 to the respective internal points are rearranged. As described above, the geodesic distance profile is preferably generated in consideration of the area of each region inside the shape data sd11.
　図６（Ｂ）は、各領域の面積を考慮した測地的距離プロファイルの一例を示す図である。図６（Ｂ）に示す測地的距離プロファイルは、各距離が小さい順に並べ替えられている。図６（Ｂ）に示すプロファイルｐｒ１１は、図６（Ａ）に示す輪郭点ａ１４の測地的距離プロファイルであり、図６（Ｂ）に示すプロファイルｐｒ１２は、図６（Ａ）に示す輪郭点ａ１２の測地的距離プロファイルであり、図６（Ｂ）に示すプロファイルｐｒ１３は、図６（Ａ）に示す輪郭点ａ１１の測地的距離プロファイルであり、図６（Ｂ）に示すプロファイルｐｒ１４は、図６（Ａ）に示す輪郭点ａ１３の測地的距離プロファイルである。 FIG. 6B is a diagram showing an example of a geodesic distance profile in consideration of the area of each region. The geodesic distance profile shown in FIG. 6B is rearranged in order of increasing distance. A profile pr11 illustrated in FIG. 6B is a geodesic distance profile of the contour point a14 illustrated in FIG. 6A, and a profile pr12 illustrated in FIG. 6B is a contour point a12 illustrated in FIG. 6A. The profile pr13 shown in FIG. 6B is a geodesic distance profile of the contour point a11 shown in FIG. 6A, and the profile pr14 shown in FIG. It is a geodesic distance profile of the outline point a13 shown to (A).
　図６（Ｂ）に示す測地的距離プロファイルを用いて、このプロファイルが持つ意味を説明する。面積（又は内部点数）の軸の値が小さい部分（図６（Ｂ）ではＬｏｃａｌと表記する。）において、距離の値が大きい場合は、輪郭点において凸になっていることを表し、Ｌｏｃａｌにおいて、距離の値が小さい場合は、輪郭点において凹になっていることを表す。これは、輪郭点において凸になっていれば、この輪郭点周辺の内部点は、この輪郭点から離れており、輪郭点において凹になっていれば、この輪郭点周辺の内部点は、この輪郭点から近いことに起因する。 The meaning of this profile will be described using the geodesic distance profile shown in FIG. In a portion where the axis value of the area (or the number of internal points) is small (indicated as “Local” in FIG. 6B), a large distance value indicates that the contour point is convex. When the distance value is small, it indicates that the contour point is concave. If the contour point is convex, the inner point around the contour point is far from the contour point. If the contour point is concave, the inner point around the contour point is This is due to being close to the contour point.
　また、面積の軸の値が大きい部分（図６（Ｂ）ではＧｌｏｂａｌと表記する。）において、距離の値が大きい場合は、輪郭点が形状データの中心（又は重心など）から離れた位置にあることを表し、Ｇｌｏｂａｌにおいて、距離の値が小さい場合は、輪郭点が形状データの中心から近い位置にあることを表す。これは、輪郭点が中心に近ければ近いほど、輪郭点が中心から離れている場合に比べて、輪郭点から一番離れた内部点までの距離が小さくなることに起因する。 Further, in a portion where the value of the area axis is large (indicated as “Global” in FIG. 6B), when the distance value is large, the contour point is located away from the center (or the center of gravity, etc.) of the shape data. When the distance value is small in Global, it indicates that the contour point is close to the center of the shape data. This is due to the fact that the closer the contour point is to the center, the smaller the distance from the contour point to the inner point that is farthest from the center as compared to the case where the contour point is far from the center.
　これにより、上述した測地的距離プロファイルは、形状データの内部領域を考慮した輪郭点の形状記述子となる。本実施形態では、この測地的距離プロファイルを用いて輪郭点のマッチングを行うことにより、従来技術では考慮されていなかった内部領域を考慮したマッチング処理を行うことが可能になる。 Thereby, the geodesic distance profile described above becomes a shape descriptor of the contour point considering the internal region of the shape data. In this embodiment, by performing contour point matching using this geodesic distance profile, it is possible to perform matching processing in consideration of an internal region that has not been considered in the prior art.
　なお、プロファイル生成部５０２は、形状データの内部領域に適切に複数の内部点を分散して設定することができれば、領域分割部５０４及び面積計算部５０８を設けなくてもよい。したがって、並べ替え部５１０において、各領域の面積に応じてリストを調整する処理は、行われなくてもよい。 Note that the profile generation unit 502 does not need to provide the region division unit 504 and the area calculation unit 508 as long as a plurality of internal points can be appropriately distributed and set in the internal region of the shape data. Therefore, the rearrangement unit 510 may not perform the process of adjusting the list according to the area of each region.
　また、プロファイル生成部５０２は、形状データ内部に閉領域が存在する場合は、閉領域の輪郭点における測地的距離プロファイルを生成する際、形状データと閉領域との間にある内部領域に内部点を設ける。これにより、形状データ内部にある閉領域に対しても、後述する図２５に示すように、適切な輪郭マッチングを行うことができる。 In addition, when a closed region exists in the shape data, the profile generation unit 502 generates an internal point in the internal region between the shape data and the closed region when generating a geodesic distance profile at the contour point of the closed region. Is provided. As a result, appropriate contour matching can be performed on the closed region inside the shape data as shown in FIG.
　図７は、実施形態におけるマッチング部２０８の構成の一例を示すブロック図である。図７に示すマッチング部２０８は、対数計算部６０２と、フィルタ処理部６０４とを含む。なお、対数計算部６０２と、フィルタ処理部６０４とは、マッチング部２０８とは別の機能として設けてもよい。 FIG. 7 is a block diagram illustrating an example of the configuration of the matching unit 208 in the embodiment. The matching unit 208 shown in FIG. 7 includes a logarithmic calculation unit 602 and a filter processing unit 604. Note that the logarithmic calculation unit 602 and the filter processing unit 604 may be provided as functions different from the matching unit 208.
　対数計算部６０２は、測地的距離プロファイルの各距離に対して対数を計算する。マッチング部２０８は、第１形状データにおいて対数が計算された各距離と、第２形状データにおいて対数が計算された各距離との差分に基づいて、第１形状データの輪郭点と、第２形状データの輪郭点とのマッチング処理を行う。 The logarithm calculation unit 602 calculates a logarithm for each distance of the geodesic distance profile. Based on the difference between each distance for which the logarithm is calculated in the first shape data and each distance for which the logarithm is calculated in the second shape data, the matching unit 208 calculates the contour points of the first shape data and the second shape Performs matching processing with contour points of data.
　対数を計算することにより、マッチング部２０８がマッチング処理を行う際、距離が短い場合の差分と比べて、距離が長い場合の差分の影響を小さくすることができる。すなわち、マッチング部２０８は、対数が計算された距離を用いてマッチング処理を行うことで、より適切なマッチング結果を得ることができる。 By calculating the logarithm, when the matching unit 208 performs the matching process, the influence of the difference when the distance is long can be reduced compared to the difference when the distance is short. That is, the matching unit 208 can obtain a more appropriate matching result by performing the matching process using the logarithm calculated distance.
　フィルタ処理部６０４は、測地的距離プロファイルに含まれる各距離に対し、フィルタ処理を実行する。フィルタ処理部６０４は、例えばローパスフィルタ又はハイパスフィルタを用いる。マッチング部２０８は、ローパスフィルタを適用することで、Ｌｏｃａｌの影響を大きくし、形状の角部分を尊重するのに適したマッチング処理を行うことができる。Ｌｏｃａｌは、人間の視覚が敏感な部分であるため、マッチング部２０８は、より直感的なマッチング処理を行うことができる。また、マッチング部２０８は、ハイパスフィルタを適用することで、Ｇｌｏｂａｌの影響を大きくし、ノイズの影響を抑圧するのに適したマッチング処理を行うことができる。 The filter processing unit 604 executes filter processing for each distance included in the geodesic distance profile. The filter processing unit 604 uses, for example, a low pass filter or a high pass filter. By applying a low-pass filter, the matching unit 208 can perform the matching process suitable for increasing the influence of Local and respecting the corners of the shape. Since Local is a part that is sensitive to human vision, the matching unit 208 can perform more intuitive matching processing. Further, the matching unit 208 can apply a high-pass filter to perform matching processing suitable for increasing the influence of Global and suppressing the influence of noise.
　フィルタ処理部６０４は、ローパスフィルタ、ハイパスフィルタのいずれを用いるかについて、用途に応じてユーザに適宜選択させるようにすればよい。 The filter processing unit 604 may make the user appropriately select whether to use a low-pass filter or a high-pass filter depending on the application.
　＜具体例＞
　次に、具体的な例を用いて、測地的距離プロファイルの生成、距離計算、及びマッチング処理までの概要について説明する。まず、図８～図１４を用いて測地的距離プロファイルの生成を説明する。 <Specific example>
Next, an outline of geodesic distance profile generation, distance calculation, and matching processing will be described using a specific example. First, generation of a geodesic distance profile will be described with reference to FIGS.
　図８は、形状データの一例を示す図である。プロファイル生成部５０２は、図８に示す形状データｓｄ２１を入力したとする。 FIG. 8 is a diagram showing an example of shape data. It is assumed that the profile generation unit 502 receives the shape data sd21 shown in FIG.
　図９は、複数の内部点の一例を示す図である。図９に示すように、プロファイル生成部５０２は、所定のアルゴリズムに従って、形状データｓｄ２１の内部に複数の内部点を設定する。このとき、プロファイル生成部５０２は、内部点同士ができるだけ近づき過ぎないように分散して設定するとよい。 FIG. 9 is a diagram illustrating an example of a plurality of internal points. As illustrated in FIG. 9, the profile generation unit 502 sets a plurality of internal points in the shape data sd21 according to a predetermined algorithm. At this time, the profile generation unit 502 may set the internal points so that they are not as close as possible.
　図１０は、輪郭点と内部点との距離の一例を示す図である。距離計算部５０６は、輪郭点ａ２１と内部点ｂ２１との距離ｄ２１、輪郭点ａ２１と内部点ｂ２２との距離ｄ２２、輪郭点ａ２１と内部点ｂ２３との距離ｄ２３、輪郭点ａ２１と内部点ｂ２４との距離ｄ２４を計算する。図１０に示すように、距離計算部５０６は、距離は、形状データｓｄ２１の内部を通る距離を計算する。これは、距離が形状データのｓｄ２１の内部領域の特徴を表すようにするためである。 FIG. 10 is a diagram illustrating an example of the distance between the contour point and the internal point. The distance calculation unit 506 includes a distance d21 between the contour point a21 and the internal point b21, a distance d22 between the contour point a21 and the internal point b22, a distance d23 between the contour point a21 and the internal point b23, and the contour point a21 and the internal point b24. The distance d24 is calculated. As illustrated in FIG. 10, the distance calculation unit 506 calculates a distance that passes through the inside of the shape data sd21. This is because the distance represents the feature of the internal region of the shape data sd21.
　図１１は、距離を並べ替えたリストの一例を示す図である。図１１に示すように、並べ替え部５１０は、距離計算部５０６により計算された各距離を、例えば小さい順に並べ替える。並べ替え部５１０により並べ替えられた各距離のリストが、測地的距離プロファイルとなる。 FIG. 11 is a diagram illustrating an example of a list in which distances are rearranged. As illustrated in FIG. 11, the rearrangement unit 510 rearranges the distances calculated by the distance calculation unit 506 in ascending order, for example. The list of distances rearranged by the rearrangement unit 510 is a geodesic distance profile.
　次に、内部領域の特徴をより反映させた測地的距離プロファイルについて説明する。図１２は、領域分割の一例を示す図である。図１２に示すように、領域分割部５０４は、形状データｓｄ２１を複数の領域ａｒ２１、ａｒ２２、ａｒ２３、ａｒ２４に分割する。領域分割部５０４は、各領域に１つの内部点が含まれるように分割する。なお、プロファイル生成部５０２は、領域分割部５０４により複数の領域に分割して内部点を設定するようにしてもよい。 Next, a geodesic distance profile that reflects the characteristics of the inner area will be described. FIG. 12 is a diagram illustrating an example of area division. As shown in FIG. 12, the area dividing unit 504 divides the shape data sd21 into a plurality of areas ar21, ar22, ar23, and ar24. The area dividing unit 504 divides each area so that one internal point is included. Note that the profile generation unit 502 may set the internal points by dividing into a plurality of regions by the region dividing unit 504.
　面積計算部５０８は、分割された各領域の面積を計算する。ここで、領域ａｒ２１の面積をｓ１、領域ａｒ２２の面積をｓ２、領域ａｒ２３の面積をｓ３、領域ａｒ２４の面積をｓ４とする。 The area calculation unit 508 calculates the area of each divided area. Here, the area of the area ar21 is s1, the area of the area ar22 is s2, the area of the area ar23 is s3, and the area of the area ar24 is s4.
　図１３は、各領域の面積を反映したリストの一例を示す図である。図１３に示すように、並べ替え部５１０は、各内部点に対し、各内部点を有する領域の面積に応じて横軸の幅を調整する。すなわち、各内部点の横軸の幅は、各内部点を有する面積の大きさに比例した長さになる。 FIG. 13 is a diagram showing an example of a list reflecting the area of each region. As shown in FIG. 13, rearrangement unit 510 adjusts the width of the horizontal axis for each internal point according to the area of the region having each internal point. That is, the width of the horizontal axis of each internal point is a length proportional to the size of the area having each internal point.
　図１４は、測地的距離プロファイルの最終結果の一例を示す図である。図１４に示すプロファイルｐｒ２１が、輪郭点ａ２１に対応付けられる形状記述子としての最終的な測地的距離プロファイルとなる。 FIG. 14 is a diagram showing an example of the final result of the geodesic distance profile. The profile pr21 shown in FIG. 14 is the final geodesic distance profile as a shape descriptor associated with the contour point a21.
　次に、距離の計算方法について説明する。図１５は、三角形メッシュ分割を用いた場合の距離の計算の一例を示す図である。図１５に示す（１）では、形状データｓｄ３１が、領域分割部５０４により三角形メッシュ分割されている。 Next, the distance calculation method will be described. FIG. 15 is a diagram illustrating an example of distance calculation in the case of using triangular mesh division. In (1) shown in FIG. 15, the shape data sd31 is divided into triangular meshes by the area dividing unit 504.
　図１５に示す（２）では、領域分割部５０４により、各三角形の中心が各内部点ｂ３１、ｂ３２、ｂ３３、ｂ３４に設定される。三角形の中心を内部点とする理由は、内部点と領域との関係が１対１で対応しやすくなるからである。 15 (2), the center of each triangle is set to each internal point b31, b32, b33, b34 by the area dividing unit 504. The reason why the center of the triangle is the internal point is that the relationship between the internal point and the region can be easily handled in a one-to-one relationship.
　図１５に示す（３）では、距離計算部５０６により、各三角形の中心から、その中心を含む三角形の各頂点及び隣接する他の三角形の中心までのパスが設定される。なお、各三角形の辺もパスとして設定される。 In (3) shown in FIG. 15, the distance calculation unit 506 sets a path from the center of each triangle to each vertex of the triangle including the center and the center of another adjacent triangle. Note that the sides of each triangle are also set as paths.
　図１５に示す（４）では、距離計算部５０６により、輪郭点ａ３１から各内部点までの最短のパスの長さが距離として計算される。 15 (4), the distance calculation unit 506 calculates the shortest path length from the contour point a31 to each internal point as the distance.
　このとき、距離計算部５０６は、輪郭点ａ３１と内部点ｂ３１との距離をパスｐｓ３１の長さとし、輪郭点ａ３１と内部点ｂ３２との距離をパスｐｓ３１とパスｐｓ３２との長さとし、輪郭点ａ３１と内部点ｂ３３との距離をパスｐｓ３３とパスｐｓ３４との長さとし、輪郭点ａ３１と内部点ｂ３４との距離をパスｐｓ３３とパスｐｓ３５との長さとする。これにより、距離計算部５０６は、形状データの内部を通る距離を比較的容易に計算することができる。 At this time, the distance calculation unit 506 sets the distance between the contour point a31 and the internal point b31 as the length of the path ps31, sets the distance between the contour point a31 and the internal point b32 as the length between the path ps31 and the path ps32, and sets the contour point a31. Is the length of the path ps33 and the path ps34, and the distance between the contour point a31 and the internal point b34 is the length of the path ps33 and the path ps35. Thereby, the distance calculation part 506 can calculate the distance which passes the inside of shape data comparatively easily.
　次に、マッチング処理までの概要について図１６を用いて説明する。図１６は、実施形態におけるマッチング処理までの概要を、イメージを用いて説明するための図である。図１６に示す（１）では、入力部２０２は、形状データｓｄ３１と、形状データｓｄ４１とを入力する。 Next, an outline up to the matching process will be described with reference to FIG. FIG. 16 is a diagram for explaining an outline of matching processing in the embodiment using an image. In (1) shown in FIG. 16, the input unit 202 inputs shape data sd31 and shape data sd41.
　図１６に示す（２）では、設定部２０６は、サンプリングされた輪郭点ごとに、測地的距離プロファイルを設定する。プロファイルｐｒ３１は、形状データｓｄ３１の輪郭点ａ３１の測地的距離プロファイルを表す。また、プロファイルｐｒ４１は、形状データｓｄ４１の輪郭点ａ４１の測地的距離プロファイルを表す。 16 (2), the setting unit 206 sets a geodesic distance profile for each sampled contour point. The profile pr31 represents a geodesic distance profile of the contour point a31 of the shape data sd31. The profile pr41 represents a geodesic distance profile of the contour point a41 of the shape data sd41.
　図１６に示す（３）では、マッチング部２０８は、測地的距離プロファイルの距離（ｄ）に対し、対数を計算する。プロファイルｐｒ３２は、対数計算後の測地的距離プロファイルを表し、プロファイルｐｒ４２は、対数計算後の測地的距離プロファイルを表す。 16 (3), the matching unit 208 calculates a logarithm with respect to the distance (d) of the geodetic distance profile. The profile pr32 represents a geodetic distance profile after logarithm calculation, and the profile pr42 represents a geodetic distance profile after logarithm calculation.
　図１６に示す（４）では、マッチング部２０８は、対数計算後の測地的距離プロファイル同士の差分を計算し、この差分を基に、２つの形状データにおける輪郭点のマッチング処理を行う。差分は、例えば二乗誤差などがある。マッチング部２０８は、１つの形状データの所定の輪郭点において、測地的距離プロファイルの差分が最も小さくなる他の形状データの輪郭点を求める。 In (4) shown in FIG. 16, the matching unit 208 calculates a difference between geodetic distance profiles after logarithmic calculation, and performs matching processing of contour points in the two shape data based on this difference. The difference includes, for example, a square error. The matching unit 208 obtains contour points of other shape data in which the difference in geodesic distance profile is the smallest at a predetermined contour point of one shape data.
　以上のマッチング処理を行うことで、形状データの内部領域を考慮した輪郭点のマッチング処理が行われる。さらに、内部点を含む領域の面積や、距離（ｄ）の対数を計算することにより、情報処理装置１０は、マッチングの精度を向上させることができる。 By performing the above matching process, the contour point matching process considering the internal region of the shape data is performed. Further, by calculating the area of the region including the internal point and the logarithm of the distance (d), the information processing apparatus 10 can improve the matching accuracy.
　＜動作＞
　次に、情報処理装置１０の動作について説明する。図１７は、実施形態における形状データの入力からマッチングまでの処理の一例を示すフローチャートである。図１７に示すステップＳ１０２で、入力部２０２は、第１形状データと第２形状データとを入力する。入力部２０２は、ソースとなる第１形状データを先に入力し、ターゲットとなる第２形状データを後に入力するようにしてもよい。 <Operation>
Next, the operation of the information processing apparatus 10 will be described. FIG. 17 is a flowchart illustrating an example of processing from input of shape data to matching in the embodiment. In step S102 shown in FIG. 17, the input unit 202 inputs the first shape data and the second shape data. The input unit 202 may input the first shape data serving as the source first and input the second shape data serving as the target later.
　ステップＳ１０４で、サンプリング部２０４は、第１形状データにおける輪郭点、及び第２形状データにおける輪郭点をサンプリングする。サンプリング部２０４は、なるべく等間隔になるように輪郭点をサンプリングする。 In step S104, the sampling unit 204 samples the contour points in the first shape data and the contour points in the second shape data. The sampling unit 204 samples the contour points so as to be as evenly spaced as possible.
　ステップＳ１０６で、設定部２０６は、第１形状データの輪郭点において第１形状記述子を設定し、第２形状データの輪郭点において第２形状記述子を設定する。ここで、形状記述子とは、例えば、上述したような形状データの内部領域を考慮した測地的距離プロファイルである。測地的距離プロファイルの生成については、図１８を用いて説明する。 In step S106, the setting unit 206 sets the first shape descriptor at the contour point of the first shape data, and sets the second shape descriptor at the contour point of the second shape data. Here, the shape descriptor is, for example, a geodetic distance profile considering the internal region of the shape data as described above. Generation of the geodesic distance profile will be described with reference to FIG.
　ステップＳ１０８で、マッチング部２０８は、第１形状記述子と第２形状記述子とを比較して、第１形状データの輪郭における各点と、第２形状データの輪郭における各点との対応付けを行う。例えば、マッチング部２０８は、第１形状データの輪郭点に対し、対数が計算された距離を有する測地的距離プロファイル同士の二乗誤差が最も小さくなる第２形状データの輪郭点を求める。マッチング部２０８は、第１形状データの全ての輪郭点に対し、マッチングする第２形状データの輪郭点を求める。 In step S108, the matching unit 208 compares the first shape descriptor with the second shape descriptor, and associates each point in the contour of the first shape data with each point in the contour of the second shape data. I do. For example, the matching unit 208 obtains the contour point of the second shape data that minimizes the square error between the geodesic distance profiles having the logarithm calculated distances with respect to the contour point of the first shape data. The matching unit 208 obtains matching contour points of the second shape data for all contour points of the first shape data.
　ステップＳ１１０で、出力部２１０は、マッチング処理により求められた第１形状データの輪郭点と、第２形状データの輪郭点との組を出力する。 In step S110, the output unit 210 outputs a set of the contour point of the first shape data and the contour point of the second shape data obtained by the matching process.
　以上の処理を行うことで、情報処理装置１０は、形状の内部領域を考慮した、形状の輪郭に対するマッチング処理を行うことができる。 By performing the above processing, the information processing apparatus 10 can perform matching processing on the contour of the shape in consideration of the internal region of the shape.
　図１８は、実施形態における測地的距離プロファイルの生成処理の一例を示すフローチャートである。図１８に示す処理は、１つの形状データでサンプリングされた全輪郭点において、測地的距離プロファイルを生成する処理を示す。 FIG. 18 is a flowchart illustrating an example of geodetic distance profile generation processing in the embodiment. The process shown in FIG. 18 shows a process of generating a geodesic distance profile at all contour points sampled with one shape data.
　ステップＳ２０２で、領域分割部５０４は、形状データを複数の領域に分割する。例えば、領域分割部５０４は、三角形メッシュ分割を用いて形状データを複数の領域に分割する。 In step S202, the region dividing unit 504 divides the shape data into a plurality of regions. For example, the region dividing unit 504 divides the shape data into a plurality of regions using triangular mesh division.
　ステップＳ２０４で、面積計算部５０８は、分割された各領域の面積を計算する。例えば、面積計算部５０８は、三角形メッシュ分割された各三角形の面積を計算する。 In step S204, the area calculation unit 508 calculates the area of each divided region. For example, the area calculation unit 508 calculates the area of each triangle divided into triangle meshes.
　ステップＳ２０６で、距離計算部５０６は、形状データの任意の１つの輪郭点と、形状データの各内部点との距離を計算する。距離計算部５０６は、形状データの内部を通る距離を計算する。 In step S206, the distance calculation unit 506 calculates the distance between any one contour point of the shape data and each internal point of the shape data. The distance calculation unit 506 calculates a distance passing through the inside of the shape data.
　ステップＳ２０８で、並べ替え部５１０は、１つの輪郭点における各距離を小さい順に並べ替えてリストを生成する。このリストが、測地的距離プロファイルとなる。 In step S208, the rearrangement unit 510 rearranges the distances at one contour point in ascending order to generate a list. This list becomes a geodesic distance profile.
　ステップＳ２１０で、並べ替え部５１０は、各領域の面積を測地的距離プロファイルに反映する。反映の仕方は、図１３を用いて説明したように、リストにおける内部点の軸の幅を、その内部点を含む領域の面積に比例して調整する。これにより、各領域の面積が反映された測地的距離プロファイルが生成される。 In step S210, the rearrangement unit 510 reflects the area of each region in the geodesic distance profile. In the reflection method, as described with reference to FIG. 13, the width of the axis of the internal point in the list is adjusted in proportion to the area of the region including the internal point. Thereby, a geodesic distance profile reflecting the area of each region is generated.
　ステップＳ２１２で、プロファイル生成部５０２は、全ての輪郭点において測地的距離プロファイルを生成したかを判定する。全輪郭点において測地的距離プロファイルが生成されていれば（ステップＳ２１２－ＹＥＳ）プロファイル生成処理が終了する。全輪郭点において測地的距離プロファイルが生成されていなければ（ステップＳ２１２－ＮＯ）別の輪郭点における測地的距離プロファイルを生成するため、ステップＳ２０６の処理が再び実行される。 In step S212, the profile generation unit 502 determines whether geodesic distance profiles have been generated at all contour points. If geodesic distance profiles have been generated at all contour points (step S212—YES), the profile generation process ends. If no geodetic distance profile has been generated at all contour points (step S212—NO), the process of step S206 is executed again to generate a geodetic distance profile at another contour point.
　以上の処理を行うことで、プロファイル生成部５０２は、形状データの内部の特徴を好適に反映する測地的距離プロファイルを生成することができる。 By performing the above processing, the profile generation unit 502 can generate a geodesic distance profile that suitably reflects the internal characteristics of the shape data.
　＜実験結果＞
　次に、本実施形態における方法と、従来の輪郭マッチングでよく用いられる非特許文献１の方法（curved based method）との実験結果を以下に示す。本実施形態の方法は、三角形メッシュ分割、各領域の面積を反映した測地的距離プロファイルを用いる。従来の方法は、輪郭の長さのパラメータと角度のパラメータとを用いてマッチングを行う。角度のパラメータには、重みｗが乗算される。なお、この重みｗは、ユーザにより適切な値が設定される必要がある。 <Experimental result>
Next, experimental results of the method according to the present embodiment and the method (curved based method) of Non-Patent Document 1 often used in conventional contour matching are shown below. The method according to the present embodiment uses a geodesic distance profile reflecting triangle mesh division and the area of each region. In the conventional method, matching is performed using a contour length parameter and an angle parameter. The angle parameter is multiplied by the weight w. The weight w needs to be set to an appropriate value by the user.
　（実験１）
　図１９は、実験１に用いた形状データの一例を示す図である。図１９（Ａ）は、ソースとなる第１形状データを示し、図１９（Ｂ）は、ターゲットとなる第２形状データを示す。図１９に示す形状データは、三角形メッシュ分割された状態の形状データである。情報処理装置１０は、第１形状データの輪郭点が、第２形状データのどの輪郭点にマッチングするかを求める。 (Experiment 1)
FIG. 19 is a diagram illustrating an example of shape data used in Experiment 1. In FIG. FIG. 19A shows the first shape data as a source, and FIG. 19B shows the second shape data as a target. The shape data shown in FIG. 19 is shape data in a state where the triangular mesh is divided. The information processing apparatus 10 determines which contour point of the second shape data matches the contour point of the first shape data.
　図２０は、実験１における輪郭マッチングを行う際の内部領域の歪み具合を示す図である。図２０（Ａ）は、第１形状データの内部領域の状態を示す。図２０（Ｂ）は、本実施形態によりマッチングされた輪郭点に基づく内部領域を示す。図２０（Ａ）と図２０（Ｂ）とを参照すると、第１形状データの頭や足などは、第２形状データにおいても適切に対応付けられている。よって、内部領域の歪の防止を図ることができる。 FIG. 20 is a diagram showing the degree of distortion in the internal region when performing contour matching in Experiment 1. FIG. FIG. 20A shows the state of the internal region of the first shape data. FIG. 20B shows an internal region based on contour points matched according to the present embodiment. Referring to FIG. 20A and FIG. 20B, the heads and feet of the first shape data are appropriately associated with each other in the second shape data. Therefore, it is possible to prevent the distortion of the inner region.
　図２０（Ｃ）は、ｗ＝１とした場合の従来方法によりマッチングされた輪郭点に基づく内部領域を示す。図２０（Ｃ）に示す矢印の部分は、本来は第１形状データの目部分の輪郭が対応付けられるはずが、実際は首部分の輪郭が対応付けられている。これは、マッチング処理が適切ではなく、このマッチング処理が用いられると内部領域が歪んでしまうことを示している。 FIG. 20C shows an internal region based on contour points matched by the conventional method when w = 1. The part indicated by the arrow shown in FIG. 20C is supposed to be associated with the outline of the eye part of the first shape data, but is actually associated with the outline of the neck part. This indicates that the matching process is not appropriate and the inner region is distorted when this matching process is used.
　図２０（Ｄ）は、ｗ＝１０とした場合の従来方法によりマッチングされた輪郭点に基づく内部領域を示す。図２０（Ｄ）に示す矢印の部分は、本来は第１形状データの前足部分の輪郭が対応付けられるはずが、実際は尾部分の輪郭が対応付けられている。これは、マッチング処理が適切ではなく、このマッチング処理が用いられると内部領域が歪んでしまうことを示している。 FIG. 20D shows an internal region based on contour points matched by the conventional method when w = 10. 20D is supposed to be associated with the contour of the forefoot portion of the first shape data, but is actually associated with the contour of the tail portion. This indicates that the matching process is not appropriate and the inner region is distorted when this matching process is used.
　図２１は、実験１における輪郭マッチングによりテクスチャトランスフォームされたテクスチャを示す図である。図２１（Ａ）は、第１形状データに動物Ａ（例えば猫）を貼り付けた状態を示す。図２１（Ｂ）は、本実施形態によりテクスチャトランスフォームされた動物Ａを示す。図２１（Ａ）と図２１（Ｂ）とを参照すると、第１形状データの輪郭は、第２形状データの輪郭において適切に対応付けられている。よって、内部領域の歪の防止を図ることができる。 FIG. 21 is a diagram showing a texture that has been texture-transformed by contour matching in Experiment 1. FIG. FIG. 21A shows a state in which an animal A (for example, a cat) is pasted on the first shape data. FIG. 21B shows animal A texture-transformed according to this embodiment. Referring to FIGS. 21A and 21B, the outline of the first shape data is appropriately associated with the outline of the second shape data. Therefore, it is possible to prevent the distortion of the inner region.
　図２１（Ｃ）は、ｗ＝１とした場合の従来方法によりテクスチャトランスフォームされた動物Ａ（例えば猫）を示す。図２１（Ｄ）は、ｗ＝１０とした場合の従来方法によりテクスチャトランスフォームされた動物Ａを示す。図２１（Ｃ）と図２１（Ｄ）とに示す矢印部分は、図２０同様、輪郭点のマッチングが適切に行われなかった箇所を示す。特に、図２１（Ｄ）に示す例では、ほとんどの輪郭点のマッチングが適切に行われていない。 FIG. 21C shows an animal A (for example, a cat) texture-transformed by a conventional method when w = 1. FIG. 21 (D) shows animal A texture-transformed by the conventional method when w = 10. The arrow part shown to FIG.21 (C) and FIG.21 (D) shows the location where matching of the contour point was not performed appropriately like FIG. In particular, in the example shown in FIG. 21D, most contour points are not appropriately matched.
　（実験２）
　図２２は、実験２に用いた形状データの一例を示す図である。図２２（Ａ）は、ソースとなる第１形状データを示し、図２２（Ｂ）は、ターゲットとなる第２形状データを示す。図２２に示す形状データは、三角形メッシュ分割された状態の形状データである。情報処理装置１０は、第１形状データの輪郭点が、第２形状データのどの輪郭点にマッチングするかを求める。 (Experiment 2)
FIG. 22 is a diagram illustrating an example of shape data used in Experiment 2. In FIG. FIG. 22A shows the first shape data as the source, and FIG. 22B shows the second shape data as the target. The shape data shown in FIG. 22 is shape data in a state where the triangular mesh is divided. The information processing apparatus 10 determines which contour point of the second shape data matches the contour point of the first shape data.
　図２３は、実験２における輪郭マッチングを行う際の内部領域の歪み具合を示す図である。図２３（Ａ）は、第１形状データの内部領域の状態を示す。図２３（Ｂ）は、本実施形態によりマッチングされた輪郭点に基づく内部領域を示す。図２３（Ａ）と図２３（Ｂ）とを参照すると、第１形状データの頭や足などは、第２形状データにおいても適切に対応付けられている。よって、内部領域の歪の防止を図ることができる。 FIG. 23 is a diagram showing the degree of distortion in the internal region when performing contour matching in Experiment 2. FIG. FIG. 23A shows the state of the internal area of the first shape data. FIG. 23B shows an internal region based on contour points matched according to the present embodiment. Referring to FIG. 23A and FIG. 23B, the heads and feet of the first shape data are appropriately associated also in the second shape data. Therefore, it is possible to prevent the distortion of the inner region.
　図２３（Ｃ）は、ｗ＝１とした場合の従来方法によりマッチングされた輪郭点に基づく内部領域を示す。図２３（Ｃ）に示す矢印の部分は、本来は第１形状データの足先部分の輪郭が対応付けられるはずが、実際は尾部分の輪郭が対応付けられている。これは、マッチング処理が適切ではなく、このマッチング処理が用いられると内部領域が歪んでしまうことを示している。 FIG. 23C shows an internal region based on contour points matched by the conventional method when w = 1. The part indicated by the arrow in FIG. 23C is supposed to be associated with the outline of the toe part of the first shape data, but is actually associated with the outline of the tail part. This indicates that the matching process is not appropriate and the inner region is distorted when this matching process is used.
　図２３（Ｄ）は、ｗ＝１０とした場合の従来方法によりマッチングされた輪郭点に基づく内部領域を示す。図２３（Ｄ）に示す矢印の部分は、本来は第１形状データの後足の一部分の輪郭が対応付けられるはずが、実際は後足の先部分の輪郭が対応付けられている。これは、マッチング処理が適切ではなく、このマッチング処理が用いられると内部領域が歪んでしまうことを示している。 FIG. 23D shows an internal region based on contour points matched by the conventional method when w = 10. The part of the arrow shown in FIG. 23D is supposed to be associated with the outline of a part of the hind legs of the first shape data, but is actually associated with the outline of the front part of the hind legs. This indicates that the matching process is not appropriate and the inner region is distorted when this matching process is used.
　図２４は、実験２における輪郭マッチングによりテクスチャトランスフォームされたテクスチャを示す図である。図２４（Ａ）は、第１形状データに動物Ｂ（例えば犬）のテクスチャを貼り付けた状態を示す。図２４（Ｂ）は、本実施形態によりテクスチャトランスフォームされた動物Ｂ）を示す。図２４（Ａ）と図２４（Ｂ）とを参照すると、第１形状データの輪郭は、第２形状データの輪郭において適切に対応付けられている。よって、内部領域の歪の防止を図ることができる。 FIG. 24 is a diagram showing a texture that has been texture-transformed by contour matching in Experiment 2. FIG. FIG. 24A shows a state where the texture of animal B (for example, dog) is pasted on the first shape data. FIG. 24B shows an animal B) that has been texture transformed according to this embodiment. Referring to FIGS. 24A and 24B, the contour of the first shape data is appropriately associated with the contour of the second shape data. Therefore, it is possible to prevent the distortion of the inner region.
　図２４（Ｃ）は、ｗ＝１とした場合の従来方法によりテクスチャトランスフォームされた動物Ｂ（例えば犬）を示す。図２４（Ｄ）は、ｗ＝１０とした場合の従来方法によりテクスチャトランスフォームされた動物Ｂを示す。図２４（Ｃ）と図２４（Ｄ）とに示す矢印部分は、図２３同様、輪郭点のマッチングが適切に行われなかった箇所を示す。 FIG. 24C shows an animal B (for example, a dog) texture-transformed by a conventional method when w = 1. FIG. 24D shows animal B texture-transformed by the conventional method when w = 10. The arrow part shown to FIG.24 (C) and FIG.24 (D) shows the location where matching of the contour point was not performed appropriately like FIG.
　上記実験１及び実験２に示すとおり、本実施形態の輪郭マッチングによれば、内部領域の歪みを防止することができる。また、従来方法では、重みによってマッチングされた輪郭点は大きく異なる。この重みは、ユーザにより入力されるため、ユーザは、何度も重みを入力し、最適な重みを見つけなければならない。一方で、本実施形態における方法では、ユーザにより設定されるパラメータはなく、自動で適切な輪郭のマッチング処理を行うことができる。 As shown in the above Experiment 1 and Experiment 2, according to the contour matching of the present embodiment, distortion of the internal region can be prevented. In the conventional method, the contour points matched by the weight are greatly different. Since this weight is input by the user, the user must input the weight many times to find the optimum weight. On the other hand, in the method according to the present embodiment, there is no parameter set by the user, and an appropriate contour matching process can be automatically performed.
　（実験３）
　次に、形状データの内部に閉領域がある場合に、本実施形態の方法と、上記従来方法との実験結果について説明する。図２５は、実験３に用いる形状データの一例を示す図である。図２５（Ａ）は、ソースとなる第１形状データを示し、図２５（Ｂ）は、ターゲットとなる第２形状データを示す。図２５に示すように、形状データ内部には、閉領域が存在する。 (Experiment 3)
Next, when there is a closed region in the shape data, the experimental results of the method of the present embodiment and the conventional method will be described. FIG. 25 is a diagram illustrating an example of shape data used in Experiment 3. FIG. 25A shows the first shape data as the source, and FIG. 25B shows the second shape data as the target. As shown in FIG. 25, a closed region exists inside the shape data.
　図２６は、実験３の実験結果を示す図である。図２６（Ａ）は、従来方法によりマッチングされた輪郭点に基づく内部領域を示す。図２５（Ａ）と図２６（Ａ）とに示すとおり、矢印が異なる部分を指すので、閉領域の輪郭マッチングは適切に行われていないことが分かる。この理由は、従来方法では、形状データの輪郭と、内部の閉領域の輪郭とが独立してマッチング処理が行われるからである。 FIG. 26 is a diagram showing the experimental results of Experiment 3. FIG. 26A shows an internal region based on contour points matched by the conventional method. As shown in FIG. 25 (A) and FIG. 26 (A), since the arrows indicate different portions, it is understood that the contour matching of the closed region is not properly performed. This is because in the conventional method, the contour of the shape data and the contour of the internal closed region are matched independently.
　図２６（Ｂ）は、本実施形態の方法によりマッチングされた輪郭点に基づく内部領域を示す。図２５（Ａ）と図２６（Ｂ）とに示すとおり、矢印はほぼ同じ部分を指すので、閉領域の輪郭マッチングは適切に行われていることが分かる。この理由は、閉領域の輪郭において、形状データと閉領域との間にある内部領域の特徴が形状記述子としてマッチングに用いられるからである。 FIG. 26B shows an internal region based on contour points matched by the method of the present embodiment. As shown in FIGS. 25 (A) and 26 (B), since the arrows indicate substantially the same part, it is understood that the contour matching of the closed region is appropriately performed. This is because, in the contour of the closed region, the feature of the internal region between the shape data and the closed region is used for matching as a shape descriptor.
　（実験４）
　次に、マッチング処理を行う際に、測地的距離プロファイルにフィルタ処理を施す場合の実験結果について説明する。図２７は、実験４の実験結果を示す図である。図２７（Ａ）は、ソースとなる形状データ及び内部領域を示す。 (Experiment 4)
Next, an experimental result in the case of performing a filtering process on a geodesic distance profile when performing a matching process will be described. FIG. 27 is a diagram illustrating experimental results of Experiment 4. FIG. 27A shows shape data as a source and internal regions.
　図２７（Ｂ）は、ターゲットとなる形状データと、フィルタ処理なしの測地的距離プロファイルを用いた場合の内部領域を示す。図２７（Ｂ）に示す矢印によれば、ソース形状データとターゲット形状データとは、左右逆の輪郭にマッチングされている。 FIG. 27 (B) shows an internal region in the case of using target shape data and a geodesic distance profile without filtering. According to the arrow shown in FIG. 27 (B), the source shape data and the target shape data are matched with the left and right outlines.
　図２７（Ｃ）は、ターゲットとなる形状データと、ローパスフィルタ処理ありの測地的距離プロファイルを用いた場合の内部領域を示す。図２７（Ｃ）に示す矢印によれば、ソースとなる形状データと、ターゲットとなる形状データとは、適切に輪郭マッチングが行われている。これは、測地的距離プロファイルにローパスフィルタを通すことで、マッチング処理において、Ｌｏｃａｌ部分が示す輪郭の凹凸の影響が強くなるからである。 FIG. 27 (C) shows an internal region in the case of using target shape data and a geodesic distance profile with low-pass filter processing. According to the arrow shown in FIG. 27C, contour matching is appropriately performed between the source shape data and the target shape data. This is because by passing a low-pass filter through the geodesic distance profile, the influence of the contour unevenness indicated by the Local portion becomes stronger in the matching process.
　以上、本実施形態によれば、形状の内部領域を考慮した、形状の輪郭に対するマッチング処理を行うことができる。また、本実施形態によれば、内部領域を考慮した形状記述子として、測地的距離プロファイルを用いることにより、内部領域を考慮した輪郭マッチング処理を適切に行うことができる。 As described above, according to the present embodiment, it is possible to perform the matching process on the contour of the shape in consideration of the internal region of the shape. Further, according to the present embodiment, the contour matching process considering the internal region can be appropriately performed by using the geodesic distance profile as the shape descriptor considering the internal region.
　なお、上記の情報処理装置１０で実行されるプログラムについて、実際のハードウェアとしては、ＣＰＵ１０２がＲＯＭ１０６からプログラムを読み出して実行することにより、上記各部のうち１又は複数の各部がＲＡＭ１０４上にロードされ、１又は複数の各部がＲＡＭ１０４上に生成されるようになっている。 As for the program executed by the information processing apparatus 10, as actual hardware, when the CPU 102 reads the program from the ROM 106 and executes it, one or more of the above-described units are loaded onto the RAM 104. One or more units are generated on the RAM 104.
　このように、上述した実施形態で説明したマッチング処理は、コンピュータに実行させるためのプログラムとして実現されてもよい。このプログラムをサーバ等からインストールしてコンピュータに実行させることで、前述したマッチング処理を実現することができる。 Thus, the matching process described in the above-described embodiment may be realized as a program for causing a computer to execute. The matching process described above can be realized by installing this program from a server or the like and causing the computer to execute it.
　また、このプログラムを記録媒体１１６に記録し、このプログラムが記録された記録媒体１１６をコンピュータに読み取らせて、前述したマッチング処理を実現させることも可能である。 It is also possible to record the program on the recording medium 116 and cause the computer to read the recording medium 116 on which the program is recorded, thereby realizing the above-described matching process.
　なお、記録媒体１１６は、ＣＤ－ＲＯＭ、フレキシブルディスク、光磁気ディスク等の様に情報を光学的，電気的或いは磁気的に記録する記録媒体、ＲＯＭ、フラッシュメモリ等の様に情報を電気的に記録する半導体メモリ等、様々なタイプの記録媒体を用いることができる。 The recording medium 116 is a recording medium that records information optically, electrically, or magnetically, such as a CD-ROM, a flexible disk, or a magneto-optical disk, and information is electrically stored such as a ROM or flash memory. Various types of recording media such as a semiconductor memory for recording can be used.
　以上、実施形態について詳述したが、上記実施形態に限定されるものではなく、特許請求の範囲に記載された範囲内において、上記実施形態以外にも種々の変形及び変更が可能である。 As mentioned above, although embodiment was explained in full detail, it is not limited to the said embodiment, A various deformation | transformation and change other than the said embodiment are possible within the range described in the claim.
　例えば、上記実施形態では、２次元の形状データを例にして説明したが、３次元の形状データについても、内部領域を考慮した形状記述子を輪郭点に設定することができる。よて、本実施形態は、３次元の形状データ同士の輪郭マッチングにも適用することができる。 For example, in the above-described embodiment, the description has been given by taking the two-dimensional shape data as an example. However, for the three-dimensional shape data, a shape descriptor considering the internal region can be set as the contour point. Therefore, this embodiment can also be applied to contour matching between three-dimensional shape data.
　また、本実施形態により処理されたマッチング結果は、アニメーションにおいて、例えば最初のフレームと最後のフレームとのキーフレームが与えられたときに、キーフレーム間のフレーム内にある形状データの補間処理に適用することができる。 In addition, the matching result processed by this embodiment is applied to the interpolation processing of shape data in the frame between the key frames when, for example, key frames of the first frame and the last frame are given in the animation. can do.
　また、このマッチング結果は、形状データの模様を移すテクスチャトランスフォームや、写真編集又は形状変形などのデフォーメーションにも適用することができる。さらに、このマッチング結果は、データベースに形状データが数多くある場合に、検索のキーとして適用することができたり、形状分類の分類識別子として適用することができたりする。 Also, this matching result can be applied to deformation such as texture transformation that transfers the pattern of shape data, photo editing, or shape deformation. Further, this matching result can be applied as a search key or a classification identifier for shape classification when there are many shape data in the database.
　また、このマッチング結果は、アニメの彩色にも適用することができる。例えば、１フレーム目の形状データに色を塗り、２フレーム目以降の形状データには、マッチング結果を用いてどこの輪郭部分にどの色を塗るかを自動的に判別することができる。 This matching result can also be applied to animation coloring. For example, the shape data of the first frame can be painted, and the shape data of the second and subsequent frames can be automatically determined which color is painted on which contour portion using the matching result.
１０　情報処理装置
１０２　ＣＰＵ
１０４　ＲＡＭ
１０６　ＲＯＭ
２０２　入力部
２０４　サンプリング部
２０６　設定部
２０８　マッチング部
２１０　出力部
４０２　拡大縮小部
５０２　プロファイル生成部
５０４　領域分割部
５０６　距離計算部
５０８　面積計算部
５１０　並べ替え部
６０２　対数計算部
６０４　フィルタ処理部 10 Information processing apparatus 102 CPU
104 RAM
106 ROM
202 Input unit 204 Sampling unit 206 Setting unit 208 Matching unit 210 Output unit 402 Enlargement / reduction unit 502 Profile generation unit 504 Area division unit 506 Distance calculation unit 508 Area calculation unit 510 Rearrangement unit 602 Logarithmic calculation unit 604 Filter processing unit

Claims (12)

  1. 　第１形状データと第２形状データとを入力する入力部と、
    　前記第１形状データ及び前記第２形状データの輪郭における各点をサンプリングするサンプリング部と、
    　前記第１形状データにおいてサンプリングされた各点に対し、前記第１形状データの内部領域に基づく第１形状記述子と、前記第２形状データにおいてサンプリングされた各点に対し、前記第２形状データの内部領域に基づく第２形状記述子とを設定する設定部と、
    　前記第１形状記述子と前記第２形状記述子とを比較して、前記第１形状データの輪郭における各点と、前記第２形状データの輪郭における各点とを対応付けるマッチング部と、
    　対応付けられた前記第１形状データの輪郭における点と、前記第２形状データの輪郭における点とのセットを出力する出力部と、
    　を備える情報処理装置。     An input unit for inputting the first shape data and the second shape data;
    A sampling unit for sampling each point in the contour of the first shape data and the second shape data;
    For each point sampled in the first shape data, a first shape descriptor based on an internal region of the first shape data, and for each point sampled in the second shape data, the second shape data A setting unit for setting a second shape descriptor based on the internal region of
    A matching unit that compares the first shape descriptor with the second shape descriptor and associates each point in the contour of the first shape data with each point in the contour of the second shape data;
    An output unit that outputs a set of the points in the contour of the first shape data and the points in the contour of the second shape data,
    An information processing apparatus comprising:
  2. 　前記設定部は、
    　前記第１形状記述子として、前記サンプリングされた点と前記第１形状データ内部の各点との、前記第１形状データ内部を通る距離を並べ替えた第１測地的距離プロファイルを生成し、
    　前記第２形状記述子として、前記サンプリングされた点と前記第２形状データ内部の各点との、前記第２形状データ内部を通る距離を並べ替えた第２測地的距離プロファイルを生成する、請求項１記載の情報処理装置。     The setting unit
    As the first shape descriptor, a first geodetic distance profile in which the distances through the first shape data between the sampled points and the respective points in the first shape data are rearranged is generated.
    Generating, as the second shape descriptor, a second geodesic distance profile in which distances between the sampled points and the points in the second shape data passing through the second shape data are rearranged. Item 6. The information processing apparatus according to Item 1.
  3. 　前記設定部は、
    　前記第１形状データ及び前記第２形状データを複数の領域に分割し、各領域内の点を前記内部の点とする、請求項２記載の情報処理装置。     The setting unit
    The information processing apparatus according to claim 2, wherein the first shape data and the second shape data are divided into a plurality of regions, and a point in each region is set as the internal point.
  4. 　前記設定部は、
    　三角形メッシュ分割を用いて、前記第１形状データ及び前記第２形状データを複数の領域に分割し、前記各領域内の点を、分割された三角形の中心点とする、請求項３記載の情報処理装置。     The setting unit
    The information according to claim 3, wherein the first shape data and the second shape data are divided into a plurality of regions by using triangular mesh division, and the points in the respective regions are set as the center points of the divided triangles. Processing equipment.
  5. 　前記設定部は、
    　前記各領域の面積を考慮して前記距離を並べ替えることにより、前記第１測地的距離プロファイル及び前記第２測地的距離プロファイルを生成する、請求項３又は４記載の情報処理装置。     The setting unit
    The information processing apparatus according to claim 3 or 4, wherein the first geodesic distance profile and the second geodesic distance profile are generated by rearranging the distances in consideration of an area of each region.
  6. 　前記マッチング部は、
    　前記第１測地的距離プロファイルの対数と前記第２測地的距離プロファイルの対数とを計算し、計算された対数同士の差分の合計が最も小さくなるように対応付け処理を行う、請求項２乃至５いずれか一項に記載の情報処理装置。     The matching unit is
    6. The logarithm of the first geodetic distance profile and the logarithm of the second geodetic distance profile are calculated, and the association process is performed so that the sum of differences between the calculated logarithms is minimized. The information processing apparatus according to any one of claims.
  7. 　前記マッチング部は、
    　前記第１測地的距離プロファイルと、前記第２測地的距離プロファイルとにフィルタ処理を行い、フィルタ処理後の各測地的距離プロファイルの差分が最も小さくなるように対応付け処理を行う、請求項２乃至５いずれか一項に記載の情報処理装置。     The matching unit is
    The first geodetic distance profile and the second geodetic distance profile are subjected to filter processing, and association processing is performed so that a difference between the geodetic distance profiles after the filter processing is minimized. The information processing apparatus according to any one of 5.
  8. 　前記第１形状データ内に第１閉領域があり、前記第２形状データ内に第２閉領域がある場合、
    　前記設定部は、
    　前記第１形状データと前記第１閉領域との間にある領域に基づいて、前記第１閉領域の輪郭における各点の第３形状記述子を設定し、
    　前記第２形状データと前記第２閉領域との間にある領域に基づいて、前記第２閉領域の輪郭における各点の第４形状記述子を設定する、請求項１記載の情報処理装置。     When there is a first closed region in the first shape data and there is a second closed region in the second shape data,
    The setting unit
    Based on the region between the first shape data and the first closed region, set a third shape descriptor of each point in the contour of the first closed region,
    The information processing apparatus according to claim 1, wherein a fourth shape descriptor of each point in the outline of the second closed region is set based on a region between the second shape data and the second closed region.
  9. 　前記サンプリング部は、
    　前記第１形状データの内部面積と前記第２形状データの内部面積とが一致するように、前記第１形状データ又は前記第２形状データを拡大又は縮小する、請求項１乃至８いずれか一項に記載の情報処理装置。     The sampling unit
    9. The first shape data or the second shape data is enlarged or reduced so that the internal area of the first shape data and the internal area of the second shape data coincide with each other. The information processing apparatus described in 1.
  10. 　コンピュータが実行する情報処理方法であって、
    　第１形状データと第２形状データとを入力するステップと、
    　前記第１形状データ及び前記第２形状データの輪郭における各点をサンプリングするステップと、
    　前記第１形状データにおいてサンプリングされた各点に対し、前記第１形状データの内部領域に基づく第１形状記述子と、前記第２形状データにおいてサンプリングされた各点に対し、前記第２形状データの内部領域に基づく第２形状記述子とを設定するステップと、
    　前記第１形状記述子と前記第２形状記述子とを比較して、前記第１形状データの輪郭における各点と、前記第２形状データの輪郭における各点とを対応付けるステップと、
    　対応付けられた前記第１形状データの輪郭における点と、前記第２形状データの輪郭における点とのセットを出力するステップと、
    　を含む情報処理方法。     An information processing method executed by a computer,
    Inputting first shape data and second shape data;
    Sampling each point in the contour of the first shape data and the second shape data;
    For each point sampled in the first shape data, a first shape descriptor based on an internal region of the first shape data, and for each point sampled in the second shape data, the second shape data Setting a second shape descriptor based on the interior region of
    Comparing the first shape descriptor with the second shape descriptor and associating each point in the contour of the first shape data with each point in the contour of the second shape data;
    Outputting a set of the points in the contour of the first shape data associated with the points in the contour of the second shape data;
    An information processing method including:
  11. 　請求項１０に記載の情報処理方法をコンピュータに実行させるためのプログラム。 A program for causing a computer to execute the information processing method according to claim 10.
  12. 　請求項１１に記載のプログラムを記録したコンピュータ読み取り可能な記録媒体。 A computer-readable recording medium on which the program according to claim 11 is recorded.
PCT/JP2014/083198 2013-12-16 2014-12-16 Information processing device, information processing method, program, and recording medium WO2015093457A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2013259116A JP2015115047A (en) 2013-12-16 2013-12-16 Information processing apparatus, information processing method, program, and recording medium
JP2013-259116 2013-12-16

Publications (1)

Publication Number Publication Date
WO2015093457A1 true WO2015093457A1 (en) 2015-06-25

Family

ID=53402803

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2014/083198 WO2015093457A1 (en) 2013-12-16 2014-12-16 Information processing device, information processing method, program, and recording medium

Country Status (2)

Country Link
JP (1) JP2015115047A (en)
WO (1) WO2015093457A1 (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6180583B1 (en) * 2016-04-28 2017-08-16 株式会社セルシス Method and program for associating respective closed regions of at least two images including a plurality of closed regions

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2009042811A (en) * 2007-08-06 2009-02-26 Univ Of Tokyo Three-dimensional shape development device, three-dimensional shape development method, and program for three-dimensional shape development

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2007134913A (en) * 2005-11-09 2007-05-31 Matsushita Electric Ind Co Ltd Method and device for selecting image
JP5290229B2 (en) * 2010-03-30 2013-09-18 セコム株式会社 Learning device and object detection device
US9013430B2 (en) * 2010-08-20 2015-04-21 University Of Massachusetts Hand and finger registration for control applications

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2009042811A (en) * 2007-08-06 2009-02-26 Univ Of Tokyo Three-dimensional shape development device, three-dimensional shape development method, and program for three-dimensional shape development

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
HIDETAKA TAKEDA ET AL.: "Similarity Retrieval and Numerical Descriptor for Shape and Color of 3D Object Models", THE INSTITUTE OF ELECTRONICS, INFORMATION AND COMMUNICATION ENGINEERS DAI 18 KAI DATA KOGAKU WORKSHOP RONBUNSHU, 1 June 2007 (2007-06-01), pages 1 - 8 *
KAZUKI OTSUKA ET AL.: "A Note on Rotation-Robust Normalization Method in 3-D Model Retrieval", IEICE TECHNICAL REPORT, vol. 112, no. 423, 24 January 2013 (2013-01-24), pages 1 - 6 *
R. OSADA ET AL.: "Shape distributions", ACM TRANSACTIONS ON GRAPHICS, vol. 21, no. 4, pages 807 - 832 *
SATORU IWAMOTO ET AL.: "Symbolic Joho o Mochiinai Gazo.Onsei Shigeki no Doji Teijisei ni Motozuku Gengo Symbol Gainen Kakutoku", IPSJ SIG NOTES CHINO SYSTEM(ICS, 15 April 2013 (2013-04-15), pages 1 - 6 *
TAKAHIRO MINAMITANI ET AL.: "Tengun no Muki to Kyori no Tokei o Mochiita 3 Jigen Model no Keijo Ruiji Hikaku", VISUAL COMPUTING GRAPHICS TO CAD GODO SYMPOSIUM 2003 YOKOSHU, 19 June 2003 (2003-06-19), pages 25 - 30 *

Also Published As

Publication number Publication date
JP2015115047A (en) 2015-06-22

Similar Documents

Publication Publication Date Title
US10424112B2 (en) Mesh boundary smoothing
US8665267B2 (en) System and method for generating 3D surface patches from unconstrained 3D curves
US20190251401A1 (en) Image composites using a generative adversarial neural network
US11704802B2 (en) Multi-dimensional model merge for style transfer
US9916684B2 (en) Generating portable three-dimensional print-preview renderings of three-dimensional models
US9865032B2 (en) Focal length warping
US20140306962A1 (en) Mesh skinning technique
US20110267347A1 (en) Systems and methods for primitive intersection in ray tracing
US8466919B1 (en) Re-rendering a portion of an image
US11450078B2 (en) Generating height maps from normal maps based on virtual boundaries
KR102137263B1 (en) Image processing apparatus and method
US11423617B2 (en) Subdividing a three-dimensional mesh utilizing a neural network
JP2022036918A (en) Uv mapping on 3d object with the use of artificial intelligence
US11501413B2 (en) Kernel reshaping-powered splatting-based efficient image space lens blur
US10083264B1 (en) Systems and methods for implicit surface modeling
WO2018039936A1 (en) Fast uv atlas generation and texture mapping
CN110192225B (en) Out-of-order pixel shader derivation
US11010939B2 (en) Rendering of cubic Bezier curves in a graphics processing unit (GPU)
JP7519462B2 (en) Method, apparatus and program for generating floorplans
US11941771B2 (en) Multi-dimensional model texture transfer
EP2618312A2 (en) System and method for snapping normals to create fair 3D surfaces from edge curves
US8289332B2 (en) Apparatus and method for determining intersections
WO2015093457A1 (en) Information processing device, information processing method, program, and recording medium
Duckworth et al. Parallel processing for real-time 3D reconstruction from video streams
US20230046431A1 (en) System and method for generating 3d objects from 2d images of garments

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 14872086

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 14872086

Country of ref document: EP

Kind code of ref document: A1