WO2013088807A1 - 症例ごとの代表画像及び読影情報を生成する方法、装置及びコンピュータプログラム - Google Patents
症例ごとの代表画像及び読影情報を生成する方法、装置及びコンピュータプログラム Download PDFInfo
- Publication number
- WO2013088807A1 WO2013088807A1 PCT/JP2012/074077 JP2012074077W WO2013088807A1 WO 2013088807 A1 WO2013088807 A1 WO 2013088807A1 JP 2012074077 W JP2012074077 W JP 2012074077W WO 2013088807 A1 WO2013088807 A1 WO 2013088807A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- calculating
- feature
- wavelet
- calculated
- Prior art date
Links
- 238000004590 computer program Methods 0.000 title claims abstract description 22
- 238000000034 method Methods 0.000 title claims abstract description 18
- 239000013598 vector Substances 0.000 claims abstract description 105
- 230000006870 function Effects 0.000 claims description 21
- 238000004364 calculation method Methods 0.000 claims description 15
- 230000000007 visual effect Effects 0.000 description 13
- 238000010586 diagram Methods 0.000 description 10
- 208000024891 symptom Diseases 0.000 description 10
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 description 6
- 230000005484 gravity Effects 0.000 description 5
- 230000009466 transformation Effects 0.000 description 5
- 201000010099 disease Diseases 0.000 description 4
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 4
- 206010041823 squamous cell carcinoma Diseases 0.000 description 4
- 238000003745 diagnosis Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 3
- 210000004072 lung Anatomy 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000000605 extraction Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000010339 dilation Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
- A61B6/5211—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data
- A61B6/5229—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data combining image data of a patient, e.g. combining a functional image with an anatomical image
- A61B6/5235—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data combining image data of a patient, e.g. combining a functional image with an anatomical image combining images from the same or different ionising radiation imaging techniques, e.g. PET and CT
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/5846—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using extracted text
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/97—Determining parameters from multiple pictures
-
- 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
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
- G06V10/446—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering using Haar-like filters, e.g. using integral image techniques
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/20—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management or administration of healthcare resources or facilities, e.g. managing hospital staff or surgery rooms
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/60—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10116—X-ray image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/03—Recognition of patterns in medical or anatomical images
Definitions
- the present invention relates to a method, an apparatus, and a computer program for generating a representative image and interpretation information for each case based on a plurality of stored images.
- Patent Document 1 discloses a similar case retrieval apparatus that retrieves similar case images and case data based on feature amounts obtained from diagnosis target images.
- the similar case search apparatus disclosed in Patent Document 1 can search for an image similar to an image obtained by imaging a patient, but usually searches for a plurality of similar images. Although they can be displayed in the order of similarity, many images with different symptoms are searched even if the image features are similar, so it may be difficult to identify the symptoms in some cases was there.
- the present invention has been made in view of such circumstances, and an object thereof is to provide a method, an apparatus, and a computer program for generating a representative image and interpretation information for each case based on a stored past image. To do.
- the method according to the first invention is a method that can be executed by a device that generates a representative image representing a case and interpretation information for each case from medical images based on past cases.
- the method according to the second aspect of the present invention calculates a two-dimensional Gabor wavelet characteristic as the wavelet characteristic in the first aspect.
- the method according to the third invention is the method according to the first or second invention, wherein M (M is a natural number of 2 or more) wavelet features are calculated for each image and binarized to calculate an M-dimensional bit string. And calculating a frequency distribution vector for all images, and calculating the spatial distance as an angle between the calculated frequency distribution vector and the centroid vector.
- an apparatus is an apparatus that generates a representative image representing a case and interpretation information for each case from medical images based on past cases.
- Wavelet feature calculating means for calculating wavelet features of a plurality of images captured and stored, keyword extracting means for extracting a keyword included in the interpretation information for each stored image, and stored images
- Information storage means for storing the calculated wavelet features and the extracted keywords in association with each other, group generation means for classifying the stored images and generating a plurality of groups based on the extracted keywords, and the generated groups Centroid vector to calculate the centroid vector of the feature vector based on the wavelet feature of each image corresponding to the keyword contained in
- a spatial distance calculating means for calculating a spatial distance between the calculating means, the calculated centroid vector, and a feature vector based on a wavelet feature of each image corresponding to the keyword included in the group; and the calculated spatial distance is the smallest
- a representative image storage unit that stores an image and interpretation information associated with the image as a representative image of
- the wavelet feature calculating means calculates a two-dimensional Gabor wavelet feature as the wavelet feature.
- an M-dimensional bit string is obtained by calculating M (M is a natural number of 2 or more) wavelet features for each image and binarizing each of the wavelet features.
- M is a natural number of 2 or more
- a frequency distribution vector calculating unit that calculates a frequency distribution vector for all images, and the spatial distance calculating unit calculates the spatial distance as an angle between the calculated frequency distribution vector and the centroid vector.
- the computer program according to the seventh invention is executed by a device that generates a representative image representing each case and interpretation information for each case from medical images based on past cases.
- a computer program capable of calculating wavelet features of a plurality of images that have been captured and stored in the past, and keywords included in the interpretation information for each stored image.
- Keyword extracting means for extracting, information storing means for storing the calculated wavelet feature and the extracted keyword in association with the stored image, a plurality of groups by classifying the stored images based on the extracted keyword Group generation means for generating the image, and the way of each image corresponding to the keyword included in the generated group
- Centroid vector calculation means for calculating a centroid vector of the feature vector based on the let feature, and calculates a spatial distance between the calculated centroid vector and the feature vector based on the wavelet feature of each image corresponding to the keyword included in the group.
- a spatial distance calculation unit, and an image having the smallest calculated spatial distance and interpretation information associated with the image are functioned as a representative image storage unit that stores the representative image of the group.
- the computer program according to the eighth invention causes the wavelet feature calculation means in the seventh invention to function as means for calculating a two-dimensional Gabor wavelet feature as the wavelet feature.
- the computer program according to a ninth aspect is the computer program according to the seventh or eighth aspect, wherein the device calculates M (M is a natural number of 2 or more) wavelet features for each image, and binarizes each.
- M is a natural number of 2 or more
- the spatial distance calculation means uses the space as the angle between the calculated frequency distribution vector and the centroid vector. It functions as a means for calculating the distance.
- the wavelet feature indicating the feature of the stored medical image is used as the feature vector
- the centroid vector is calculated as the feature vector for each case
- the feature vector having the shortest spatial distance from the centroid vector is included. Since images are stored as representative images, images showing typical cases can be used as guideline for cases, and patients can be diagnosed with a certain quality without depending on the experience and skill of doctors. Become.
- the present invention is implemented as a computer program that can be partially executed by a computer. be able to. Therefore, the present invention provides an embodiment as hardware, an embodiment as software, or software as a case image generation device that generates a representative image and interpretation information for each case from medical images based on past cases. Embodiments of combinations of hardware and hardware can be taken.
- the computer program can be recorded on any computer-readable recording medium such as a hard disk, DVD, CD, optical storage device, magnetic storage device or the like.
- the wavelet feature indicating the feature of the stored medical image is used as the feature vector
- the center of gravity vector is calculated as the feature vector for each case
- the spatial distance between the center of gravity vector is the shortest Since an image having a feature vector is stored as a representative image, an image showing a typical case can be used as a guideline for a case, and a patient is diagnosed with a certain quality without depending on the experience and skill of a doctor. It becomes possible.
- FIG. 1 is a block diagram schematically showing a configuration of a case image generation apparatus according to an embodiment of the present invention.
- the case image generation apparatus 1 includes at least a CPU (central processing unit) 11, a memory 12, a storage device 13, an I / O interface 14, a video interface 15, a portable disk drive 16, and a communication interface 17. And an internal bus 18 for connecting the hardware described above.
- the CPU 11 is connected to the above-described hardware units of the case image generation apparatus 1 via the internal bus 18, controls the operation of the above-described hardware units, and stores the computer program 100 stored in the storage device 13. Various software functions are executed according to the above.
- the memory 12 is composed of a volatile memory such as SRAM or SDRAM, and a load module is expanded when the computer program 100 is executed, and stores temporary data generated when the computer program 100 is executed.
- the storage device 13 includes a built-in fixed storage device (hard disk), a ROM, and the like.
- the computer program 100 stored in the storage device 13 is downloaded by the portable disk drive 16 from a portable recording medium 90 such as a DVD or CD-ROM in which information such as programs and data is recorded, and from the storage device 13 at the time of execution.
- the program is expanded into the memory 12 and executed.
- a computer program downloaded from an external computer connected via the communication interface 17 may be used.
- the storage device 13 includes a medical image storage unit 131, an interpretation information storage unit 132, a visual word storage unit 133, a frequency distribution information storage unit 134, and a case image database 135.
- the medical image storage unit 131 stores past image data obtained by X-ray imaging in association with identification information for identifying interpretation information.
- the interpretation information storage unit 132 stores the results of a doctor interpreting and diagnosing past medical images. For example, a doctor's diagnosis result such as “a nodule shadow is observed in the upper lobe of the left lung field. Suspected squamous cell carcinoma. Instructing detailed examination by HR-CT” is stored as text data in association with identification information.
- the visual word storage unit 133 stores a Gabor wavelet feature group described later as a visual word.
- the frequency distribution information storage unit 134 stores, as a feature vector, a frequency distribution vector having a value obtained by binarizing the calculated wavelet feature and converting it into an M-dimensional bit string.
- the case image database 135 stores a representative image, which is the most typical image for each case, and interpretation information corresponding to the representative image in a database.
- the case image database 135 functions as a guideline for cases and can extract typical images for each case. Therefore, the patient image database 135 can diagnose a patient with a certain quality without depending on the experience and skill of a doctor. Is possible.
- the communication interface 17 is connected to an internal bus 18 and can transmit / receive data to / from an external computer or the like by connecting to an external network such as the Internet, a LAN, or a WAN.
- the I / O interface 14 is connected to input devices such as a keyboard 21 and a mouse 22 and receives data input.
- the video interface 15 is connected to a display device 23 such as a CRT display or a liquid crystal display, and displays a representative image and interpretation information corresponding to the representative image on the display device 23.
- FIG. 2 is a functional block diagram of the case image generation apparatus 1 according to the embodiment of the present invention.
- a wavelet feature calculation unit 201 of the case image generation device 1 calculates wavelet features of a plurality of images that have been captured and stored in the past.
- a Gabor wavelet feature is calculated as the wavelet feature.
- FIG. 3 is an explanatory diagram of the coordinate setting in the image of the case image generating apparatus 1 according to the embodiment of the present invention.
- an image composed of m pixels in the x direction and n pixels in the y direction is defined as s (x, y) with the upper left corner of the image as the origin.
- Any i (i is a natural number) represents the th coordinate of the pixel P i P i (x i, y i) and.
- the matrix A is a 3 ⁇ 3 affine transformation matrix.
- the affine transformation that moves the entire image in the x direction by tx and the ty in the y direction can be expressed by (Equation 2)
- the affine transformation that rotates the entire image by the rotation angle ⁇ can be expressed by (Equation 3).
- the two-dimensional Gabor wavelet function is defined as shown in (Formula 4) with respect to the coordinate values (x dots, y dots) after rotating affine transformation.
- the two-dimensional Gabor wavelet function is composed of a real part and an imaginary part.
- FIG. 4 is an illustration of a two-dimensional Gabor wavelet function.
- FIG. 4A shows an example of the real part of the two-dimensional Gabor wavelet function
- FIG. 4B shows an example of the imaginary part of the two-dimensional Gabor wavelet function.
- u 0 indicates the frequency of the wave shape
- ⁇ indicates the width of the hat-like width.
- r has shown the direction mentioned later.
- the window function g sigma shown in (Equation 4) is a two-dimensional Gaussian function can be expressed by (Equation 5).
- the Gabor wavelet feature for the acquired image s (x, y) can be calculated by (Equation 6).
- the lattice point having the maximum absolute value of the Gabor wavelet feature and the Gabor wavelet feature near the lattice point are invariable even when the image is subjected to affine transformation such as enlargement / reduction or rotation. Therefore, it is suitable as a feature amount of an image.
- a j and a ⁇ j indicate parameters indicating the degree of dilation (enlargement / reduction), and x 0 and y 0 indicate parallel movement. Further, r indicates a direction, and in this embodiment, Gabor wavelet features are calculated for each of the eight directions.
- FIG. 5 is a schematic diagram showing the direction of the two-dimensional Gabor wavelet function of the case image generation apparatus 1 according to the embodiment of the present invention.
- Gabor wavelet characteristics are calculated in the directions (1) to (8), that is, in eight directions rotated by 22.5 degrees from a predetermined direction.
- the Gabor wavelet feature for example, it is possible to calculate the wavelet feature amount that absorbs the variation in the shape of the human organ, so that a more appropriate representative image can be selected from the images related to the same case. .
- the scale is a value for distinguishing the size to be enlarged / reduced, and indicates that the scale is enlarged from 1 to 5, for example.
- Gabor wavelet features having an absolute value equal to or greater than a predetermined threshold are extracted, and Gabor wavelet features having a maximum value are selected from them.
- the absolute value of the Gabor wavelet feature is a maximum value
- the absolute value of the integral value in (Equation 6) is a maximum value, and changes the average luminance of the image, or changes the scale of the image. Even if an operation such as rotating an image is performed, the feature amount remains unchanged.
- Gabor wavelet features there are 24 (3 scale ⁇ 8 directions) Gabor wavelet features, consisting of 8 Gabor wavelet features on the maximum scale and 8 Gabor wavelet features on the preceding and following scales. are stored in the visual word storage unit 133 as a set of visual words.
- FIG. 6 is an exemplary diagram of a data structure of a visual word stored in the visual word storage unit 133 of the storage device 13 of the case image generation device 1 according to the embodiment of the present invention.
- the calculated 24 Gabor wavelet features are listed and stored for each identification number 1, 2, 3,. That is, the first ‘1’ is an identification number, and the numerical values described after ‘1:’ to ‘24: ’across the blank indicate 24 calculated Gabor wavelet features.
- the example of FIG. 6 shows a visual word when three maximum values exist in one image. Therefore, in FIG. 6, visual words are stored for three identification numbers “1”, “2”, and “3”, but if the number of local maximum values is one, only the identification number “1” is stored. Needless to say.
- the keyword extracting unit 202 stores the image interpretation stored in the image interpretation information storage unit 132 of the storage device 13 corresponding to the past image stored in the medical image storage unit 131 of the storage device 13. Extract keywords included in the information. For example, if the interpretation information storage unit 132 of the storage device 13 stores “nodular shadow in the upper lobe of the left lung field. Suspected squamous cell carcinoma. Instructed detailed examination by HR-CT”. Is used for parsing and segmented into “part”, “symptom”, “disease name”, “treatment”, etc., and keywords are extracted.
- FIG. 7 is an exemplary diagram of keyword extraction of the case image generation device 1 according to the embodiment of the present invention.
- the nodal shadow is observed in the upper lobe of the left lung field.
- the suspected squamous cell carcinoma is extracted from the keyword “nodule shadow”
- "disease name” is “suspected squamous cell carcinoma”
- “treatment” is "detailed examination by HR-CT”.
- the information storage unit 203 associates the wavelet feature calculated by the above-described method and the extracted keyword with the past image stored in the medical image storage unit 131 of the storage device 13. Is stored in the visual word storage unit 133.
- the group generation unit 204 classifies the stored images based on the extracted keywords to generate a plurality of groups.
- the unit to be classified is not particularly limited, but it is preferable to generate a plurality of groups by classifying according to “part”, “symptom”, “disease name”, etc. which are items at the time of parsing, or a combination thereof. .
- one image may be classified so as to be included in a plurality of groups.
- the frequency distribution vector calculation unit 208 generates a histogram indicating the frequency distribution of the converted 24-dimensional bit string values.
- the histogram is generated for all the images included in the group.
- FIG. 8 is an exemplary diagram of a histogram according to the embodiment of the present invention.
- the horizontal axis takes the value of 2 24, obtains the frequency distribution of the respective values. Then, the frequency distribution for each image is stored in the frequency distribution information storage unit 134 of the storage device 13 as a frequency distribution vector.
- the centroid vector calculation unit 205 calculates, for each group, the centroid vector of the frequency distribution vector according to (Equation 7) using the wavelet features of the images included in the group and the calculated frequency distribution vector. That is, the centroid vector V T for the frequency distribution vector V i (i is the number of images included in the group), by dividing the sum of the frequency distribution vector V i the sum of the norm (length) of the frequency distribution vector V i Ask.
- the spatial distance calculation unit 206 calculates the spatial distance between the calculated centroid vector and the feature vector (frequency distribution vector) based on the wavelet feature of each image corresponding to the keyword included in the group. .
- the spatial distance calculation unit 206 calculates the spatial distance as an angle between the feature vector (frequency distribution vector) and the centroid vector. Specifically, when the frequency distribution vector of the images included in the group is V i and the center of gravity vector is V T , it is calculated as cos ⁇ with respect to the angle ⁇ formed by the two vectors according to (Equation 8).
- Equation 8 ⁇ V i , V T > is the inner product of the vector V i and the vector V T, and the denominator is the product of the norm (length) of the vector V i and the norm of the vector V T. , Respectively.
- the representative image storage unit 207 stores the image with the shortest calculated spatial distance and the interpretation information associated with the image in the case image database 135 as a representative image of the group.
- FIG. 9 is an exemplary view of a representative image display screen of the case image generation device 1 according to the embodiment of the present invention.
- a past image determined to be most similar to the center of gravity vector is displayed as a representative image, and a wavelet feature larger than a predetermined value among the wavelet features is superimposed on the image as a feature vector.
- the length of the arrow indicates the size of the feature amount, and the direction indicates the direction with the largest feature amount among the eight directions.
- the scale may be distinguished by the color, line type, or the like.
- FIG. 10 is a flowchart showing a processing procedure of the CPU 11 of the case image generating apparatus 1 according to the embodiment of the present invention.
- the CPU 11 of the case image generation apparatus 1 calculates wavelet features of a plurality of images that have been captured and stored in the past (step S1001).
- a Gabor wavelet feature is calculated as the wavelet feature.
- the CPU 11 extracts a keyword included in the stored interpretation information for each past image (step S1002), correlates the stored past image with the wavelet feature calculated by the method described above and the extracted keyword.
- the CPU 11 classifies the stored images based on the extracted keywords to generate a plurality of groups (step S1003).
- the unit to be classified is not particularly limited, but it is preferable to generate a plurality of groups by classifying according to “part”, “symptom”, “disease name”, etc., which are items at the time of parsing, or a combination thereof. .
- one image may be classified so as to be included in a plurality of groups.
- the CPU 11 calculates a centroid vector as a feature vector for each group based on the wavelet features of the images included in the group (step S1004).
- the CPU 11 calculates a spatial distance between the calculated centroid vector and the feature vector (frequency distribution vector) based on the wavelet feature of each image corresponding to the keyword included in the group (step S1005).
- the CPU 11 sets the minimum value to a predetermined value (step S1006), and selects one image from the images included in the group (step S1007).
- the CPU 11 determines whether or not the spatial distance calculated for the selected image is smaller than the minimum value (step S1008).
- step S1008: YES When the CPU 11 determines that it is smaller than the minimum value (step S1008: YES), the CPU 11 stores the spatial distance as the minimum value (step S1009). When the CPU 11 determines that the value is equal to or greater than the minimum value (step S1008: NO), the CPU 11 skips step S1009. The CPU 11 determines whether or not all images have been selected (step S1010). If the CPU 11 determines that there is an image that has not yet been selected (step S1010: NO), the CPU 11 selects the next image. (Step S1011), the process returns to Step S1008, and the above-described process is repeated.
- step S1010 determines that all images have been selected (step S1010: YES)
- the CPU 11 uses the image corresponding to the spatial distance stored as the minimum value and the corresponding interpretation information as the case image database 135. (Step S1012).
- a wavelet feature indicating a feature of a stored medical image is used as a feature vector, a centroid vector is calculated as a feature vector for each case, and a spatial distance from the centroid vector is calculated. Since the image with the shortest feature vector is stored as a representative image, an image showing a typical case can be used as a guideline for the case, and the patient can be maintained at a certain quality without being influenced by the experience and skill of the doctor. Diagnosis is possible.
- Case Image Generation Device CPU DESCRIPTION OF SYMBOLS 12 Memory 13 Storage device 14 I / O interface 15 Video interface 16 Portable disk drive 17 Communication interface 18 Internal bus 90 Portable recording medium 100 Computer program 131 Medical image storage part 132 Interpretation information storage part 133 Visual word storage part 134 Frequency distribution Information storage unit 135 Case image database
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Primary Health Care (AREA)
- Epidemiology (AREA)
- General Physics & Mathematics (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Pathology (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Surgery (AREA)
- High Energy & Nuclear Physics (AREA)
- Veterinary Medicine (AREA)
- Optics & Photonics (AREA)
- Animal Behavior & Ethology (AREA)
- Heart & Thoracic Surgery (AREA)
- Molecular Biology (AREA)
- General Business, Economics & Management (AREA)
- Business, Economics & Management (AREA)
- Databases & Information Systems (AREA)
- Multimedia (AREA)
- Library & Information Science (AREA)
- Quality & Reliability (AREA)
- General Engineering & Computer Science (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Medical Treatment And Welfare Office Work (AREA)
- Image Processing (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
Description
11 CPU
12 メモリ
13 記憶装置
14 I/Oインタフェース
15 ビデオインタフェース
16 可搬型ディスクドライブ
17 通信インタフェース
18 内部バス
90 可搬型記録媒体
100 コンピュータプログラム
131 医用画像記憶部
132 読影情報記憶部
133 ビジュアルワード記憶部
134 度数分布情報記憶部
135 症例画像データベース
Claims (9)
- 過去の症例に基づく医用画像の中から、症例ごとに症例を代表する代表画像及び読影情報を生成する装置で実行することが可能な方法であって、
過去に撮像して記憶してある複数の画像のウェーブレット特徴を算出するステップと、
記憶してある画像ごとの読影情報に含まれるキーワードを抽出するステップと、
記憶してある画像に対応付けて、算出したウェーブレット特徴及び抽出したキーワードを記憶するステップと、
抽出したキーワードに基づいて、記憶してある画像を分類して複数のグループを生成するステップと、
生成したグループに含まれるキーワードに対応する各画像のウェーブレット特徴に基づく特徴ベクトルの重心ベクトルを算出するステップと、
算出した重心ベクトルと、前記グループに含まれるキーワードに対応する各画像のウェーブレット特徴に基づく特徴ベクトルとの間の空間距離を算出するステップと、
算出した空間距離が最も小さい画像及び該画像に対応付けられている読影情報を前記グループの代表画像として記憶するステップと
を含む方法。 - 前記ウェーブレット特徴として、2次元のガボールウェーブレット特徴を算出する請求項1に記載の方法。
- 画像ごとにM(Mは2以上の自然数)個の前記ウェーブレット特徴を算出し、それぞれ二値化することによりM次元のビット列に換算し、すべての画像について度数分布ベクトルを算出するステップを含み、
算出した度数分布ベクトルと前記重心ベクトルとの角度として前記空間距離を算出する請求項1又は2に記載の方法。 - 過去の症例に基づく医用画像の中から、症例ごとに症例を代表する代表画像及び読影情報を生成する装置であって、
過去に撮像して記憶してある複数の画像のウェーブレット特徴を算出するウェーブレット特徴算出手段と、
記憶してある画像ごとの読影情報に含まれるキーワードを抽出するキーワード抽出手段と、
記憶してある画像に対応付けて、算出したウェーブレット特徴及び抽出したキーワードを記憶する情報記憶手段と、
抽出したキーワードに基づいて、記憶してある画像を分類して複数のグループを生成するグループ生成手段と、
生成したグループに含まれるキーワードに対応する各画像のウェーブレット特徴に基づく特徴ベクトルの重心ベクトルを算出する重心ベクトル算出手段と、
算出した重心ベクトルと、前記グループに含まれるキーワードに対応する各画像のウェーブレット特徴に基づく特徴ベクトルとの間の空間距離を算出する空間距離算出手段と、 算出した空間距離が最も小さい画像及び該画像に対応付けられている読影情報を前記グループの代表画像として記憶する代表画像記憶手段と
を備える装置。 - 前記ウェーブレット特徴算出手段は、前記ウェーブレット特徴として、2次元のガボールウェーブレット特徴を算出する請求項4に記載の装置。
- 画像ごとにM(Mは2以上の自然数)個の前記ウェーブレット特徴を算出し、それぞれ二値化することによりM次元のビット列に換算し、すべての画像について度数分布ベクトルを算出する度数分布ベクトル算出手段を備え、
前記空間距離算出手段は、算出した度数分布ベクトルと前記重心ベクトルとの角度として前記空間距離を算出する請求項4又は5に記載の装置。 - 過去の症例に基づく医用画像の中から、症例ごとに症例を代表する代表画像及び読影情報を生成する装置で実行することが可能なコンピュータプログラムであって、
前記装置を、
過去に撮像して記憶してある複数の画像のウェーブレット特徴を算出するウェーブレット特徴算出手段、
記憶してある画像ごとの読影情報に含まれるキーワードを抽出するキーワード抽出手段、
記憶してある画像に対応付けて、算出したウェーブレット特徴及び抽出したキーワードを記憶する情報記憶手段、
抽出したキーワードに基づいて、記憶してある画像を分類して複数のグループを生成するグループ生成手段、
生成したグループに含まれるキーワードに対応する各画像のウェーブレット特徴に基づく特徴ベクトルの重心ベクトルを算出する重心ベクトル算出手段、
算出した重心ベクトルと、前記グループに含まれるキーワードに対応する各画像のウェーブレット特徴に基づく特徴ベクトルとの間の空間距離を算出する空間距離算出手段、及び
算出した空間距離が最も小さい画像及び該画像に対応付けられている読影情報を前記グループの代表画像として記憶する代表画像記憶手段
として機能させるコンピュータプログラム。 - 前記ウェーブレット特徴算出手段を、前記ウェーブレット特徴として、2次元のガボールウェーブレット特徴を算出する手段として機能させる請求項7に記載のコンピュータプログラム。
- 前記装置を、画像ごとにM(Mは2以上の自然数)個の前記ウェーブレット特徴を算出し、それぞれ二値化することによりM次元のビット列に換算し、すべての画像について度数分布ベクトルを算出する度数分布ベクトル算出手段として機能させ、
前記空間距離算出手段を、算出した度数分布ベクトルと前記重心ベクトルとの角度として前記空間距離を算出する手段として機能させる請求項7又は8に記載のコンピュータプログラム。
Priority Applications (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201280061802.2A CN103988205B (zh) | 2011-12-13 | 2012-09-20 | 生成病例代表图像和射线照相图像解译信息的方法、设备 |
DE112012004826.1T DE112012004826B4 (de) | 2011-12-13 | 2012-09-20 | Verfahren, Vorrichtung und Computerprogramm zum Erzeugen eines repräsentativen Bildes und radiologischer Auswertungsdaten für jeden Fall |
JP2013549143A JP5764220B2 (ja) | 2011-12-13 | 2012-09-20 | 症例ごとの代表画像及び読影情報を生成する方法、装置及びコンピュータプログラム |
GB1410473.1A GB2515398B (en) | 2011-12-13 | 2012-09-20 | Method,apparatus, and computer program for generating representative image and radiographic interpretation information for each case |
US14/365,337 US9201901B2 (en) | 2011-12-13 | 2012-09-20 | Techniques for generating a representative image and radiographic interpretation information for a case |
US14/491,230 US9262442B2 (en) | 2012-09-20 | 2014-09-19 | Techniques for generating a representative image and radiographic interpretation information for a case |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2011-272175 | 2011-12-13 | ||
JP2011272175 | 2011-12-13 |
Related Child Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US14/365,337 A-371-Of-International US9201901B2 (en) | 2011-12-13 | 2012-09-20 | Techniques for generating a representative image and radiographic interpretation information for a case |
US14/491,230 Continuation US9262442B2 (en) | 2012-09-20 | 2014-09-19 | Techniques for generating a representative image and radiographic interpretation information for a case |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2013088807A1 true WO2013088807A1 (ja) | 2013-06-20 |
Family
ID=48612266
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2012/074077 WO2013088807A1 (ja) | 2011-12-13 | 2012-09-20 | 症例ごとの代表画像及び読影情報を生成する方法、装置及びコンピュータプログラム |
Country Status (6)
Country | Link |
---|---|
US (1) | US9201901B2 (ja) |
JP (1) | JP5764220B2 (ja) |
CN (1) | CN103988205B (ja) |
DE (1) | DE112012004826B4 (ja) |
GB (1) | GB2515398B (ja) |
WO (1) | WO2013088807A1 (ja) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP5868536B1 (ja) * | 2015-03-24 | 2016-02-24 | 株式会社テクノプロジェクト | 記録情報管理システム及び記録情報管理プログラム、並びに記録情報管理方法 |
WO2016142990A1 (ja) * | 2015-03-06 | 2016-09-15 | 富士通株式会社 | 検索プログラム、検索方法および検索装置 |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108542381B (zh) * | 2018-04-11 | 2021-08-31 | 京东方科技集团股份有限公司 | 一种数据处理方法和装置 |
US11915192B2 (en) | 2019-08-12 | 2024-02-27 | Walmart Apollo, Llc | Systems, devices, and methods for scanning a shopping space |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002132831A (ja) * | 2000-10-23 | 2002-05-10 | Olympus Optical Co Ltd | 画像表示方法および画像表示装置 |
JP2004086538A (ja) * | 2002-08-27 | 2004-03-18 | Fuji Photo Film Co Ltd | 画像の特徴量分類方法および装置 |
JP2004254742A (ja) * | 2003-02-24 | 2004-09-16 | Konica Minolta Holdings Inc | 医用画像処理装置及び悪性度の判定方法 |
JP2009297365A (ja) * | 2008-06-16 | 2009-12-24 | Olympus Corp | 画像処理装置、画像処理方法および画像処理プログラム |
JP2010079568A (ja) * | 2008-09-25 | 2010-04-08 | Canon Inc | データ検索装置およびその制御方法、データ検索システム |
JP2011040050A (ja) * | 2009-07-15 | 2011-02-24 | Nikon Corp | 画像分類装置 |
Family Cites Families (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH10326286A (ja) | 1997-05-27 | 1998-12-08 | Mitsubishi Electric Corp | 類似検索装置及び類似検索プログラムを記録した記録媒体 |
US7099860B1 (en) * | 2000-10-30 | 2006-08-29 | Microsoft Corporation | Image retrieval systems and methods with semantic and feature based relevance feedback |
CN1445696A (zh) * | 2002-03-18 | 2003-10-01 | 朗迅科技公司 | 自动检索图像数据库中相似图象的方法 |
JP3910997B2 (ja) * | 2003-07-02 | 2007-04-25 | 聰 山竹 | 画像データベースシステム |
US7734107B2 (en) * | 2006-02-24 | 2010-06-08 | Sony Corporation | System and method for performing wavelet-based texture feature extraction and classification |
WO2007105444A1 (ja) * | 2006-03-07 | 2007-09-20 | Konica Minolta Medical & Graphic, Inc. | 医用画像管理システム |
JP4799251B2 (ja) | 2006-04-05 | 2011-10-26 | 富士フイルム株式会社 | 類似症例検索装置、類似症例検索方法およびそのプログラム |
JP4905967B2 (ja) * | 2007-03-02 | 2012-03-28 | 富士フイルム株式会社 | 類似症例検索装置、方法、およびプログラム |
JP2008253292A (ja) * | 2007-03-30 | 2008-10-23 | Fujifilm Corp | 症例画像検索装置及びシステム |
JP5098559B2 (ja) * | 2007-10-11 | 2012-12-12 | 富士ゼロックス株式会社 | 類似画像検索装置、及び類似画像検索プログラム |
CN101373479A (zh) * | 2008-09-27 | 2009-02-25 | 华中科技大学 | 一种乳腺x线摄片计算机图像检索方法及*** |
CN101799806A (zh) * | 2009-02-06 | 2010-08-11 | 沈阳工大普日软件技术有限公司 | 基于多信息融合的胸部数字化x线图像检索*** |
JP5599572B2 (ja) * | 2009-03-12 | 2014-10-01 | 富士フイルム株式会社 | 症例画像検索装置、方法およびプログラム |
JP5670079B2 (ja) * | 2009-09-30 | 2015-02-18 | 富士フイルム株式会社 | 医用画像表示装置および方法、並びにプログラム |
JP5383431B2 (ja) * | 2009-10-27 | 2014-01-08 | キヤノン株式会社 | 情報処理装置、情報処理方法及びプログラム |
-
2012
- 2012-09-20 US US14/365,337 patent/US9201901B2/en active Active
- 2012-09-20 DE DE112012004826.1T patent/DE112012004826B4/de active Active
- 2012-09-20 CN CN201280061802.2A patent/CN103988205B/zh active Active
- 2012-09-20 GB GB1410473.1A patent/GB2515398B/en active Active
- 2012-09-20 WO PCT/JP2012/074077 patent/WO2013088807A1/ja active Application Filing
- 2012-09-20 JP JP2013549143A patent/JP5764220B2/ja active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002132831A (ja) * | 2000-10-23 | 2002-05-10 | Olympus Optical Co Ltd | 画像表示方法および画像表示装置 |
JP2004086538A (ja) * | 2002-08-27 | 2004-03-18 | Fuji Photo Film Co Ltd | 画像の特徴量分類方法および装置 |
JP2004254742A (ja) * | 2003-02-24 | 2004-09-16 | Konica Minolta Holdings Inc | 医用画像処理装置及び悪性度の判定方法 |
JP2009297365A (ja) * | 2008-06-16 | 2009-12-24 | Olympus Corp | 画像処理装置、画像処理方法および画像処理プログラム |
JP2010079568A (ja) * | 2008-09-25 | 2010-04-08 | Canon Inc | データ検索装置およびその制御方法、データ検索システム |
JP2011040050A (ja) * | 2009-07-15 | 2011-02-24 | Nikon Corp | 画像分類装置 |
Non-Patent Citations (1)
Title |
---|
AI GOMI ET AL.: "CAT : A Graphical User Interface for Visualization and Level-of-Detail Control for Large Scale Image Collections", THE JOURNAL OF THE INSTITUTE OF IMAGE ELECTRONICS ENGINEERS OF JAPAN, vol. 37, no. 4, 25 July 2008 (2008-07-25), pages 436 - 443 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2016142990A1 (ja) * | 2015-03-06 | 2016-09-15 | 富士通株式会社 | 検索プログラム、検索方法および検索装置 |
JPWO2016142990A1 (ja) * | 2015-03-06 | 2017-11-30 | 富士通株式会社 | 検索プログラム、検索方法および検索装置 |
JP5868536B1 (ja) * | 2015-03-24 | 2016-02-24 | 株式会社テクノプロジェクト | 記録情報管理システム及び記録情報管理プログラム、並びに記録情報管理方法 |
Also Published As
Publication number | Publication date |
---|---|
JPWO2013088807A1 (ja) | 2015-04-27 |
GB2515398B (en) | 2018-02-14 |
US9201901B2 (en) | 2015-12-01 |
GB2515398A (en) | 2014-12-24 |
CN103988205B (zh) | 2017-02-08 |
DE112012004826B4 (de) | 2018-05-30 |
CN103988205A (zh) | 2014-08-13 |
GB201410473D0 (en) | 2014-07-30 |
US20140307938A1 (en) | 2014-10-16 |
DE112012004826T5 (de) | 2014-08-14 |
JP5764220B2 (ja) | 2015-08-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10452899B2 (en) | Unsupervised deep representation learning for fine-grained body part recognition | |
US20220036564A1 (en) | Method of classifying lesion of chest x-ray radiograph based on data normalization and local patch and apparatus thereof | |
Guan et al. | Thigh fracture detection using deep learning method based on new dilated convolutional feature pyramid network | |
US8958614B2 (en) | Image-based detection using hierarchical learning | |
Segal et al. | Evaluating the clinical realism of synthetic chest x-rays generated using progressively growing gans | |
CN111095426A (zh) | 使用深度神经网络的计算机辅助诊断 | |
WO2018176035A1 (en) | Method and system of building hospital-scale chest x-ray database for entity extraction and weakly-supervised classification and localization of common thorax diseases | |
JP5902196B2 (ja) | 医用画像を検索する方法、装置及びコンピュータプログラム | |
JP2014029644A (ja) | 類似症例検索装置および類似症例検索方法 | |
Hassan et al. | COVID-19 diagnosis-based deep learning approaches for COVIDx dataset: A preliminary survey | |
JP5764220B2 (ja) | 症例ごとの代表画像及び読影情報を生成する方法、装置及びコンピュータプログラム | |
Pham et al. | Chest X-ray abnormalities localization via ensemble of deep convolutional neural networks | |
Pham et al. | Chest x-rays abnormalities localization and classification using an ensemble framework of deep convolutional neural networks | |
Wali et al. | Stynmedgan: medical images augmentation using a new GAN model for improved diagnosis of diseases | |
US9262442B2 (en) | Techniques for generating a representative image and radiographic interpretation information for a case | |
Wang et al. | U-shaped GAN for semi-supervised learning and unsupervised domain adaptation in high resolution chest radiograph segmentation | |
Chamveha et al. | Local adaptation improves accuracy of deep learning model for automated x-ray thoracic disease detection: A thai study | |
Moghaddam et al. | Towards smart diagnostic methods for COVID-19: Review of deep learning for medical imaging | |
EP4327333A1 (en) | Methods and systems for automated follow-up reading of medical image data | |
Basiri et al. | Synthesizing diabetic foot ulcer images with diffusion model | |
Kovalev et al. | Automatic detection of pathological changes in chest X-ray screening images using deep learning methods | |
Keshavamurthy et al. | Weakly supervised pneumonia localization in chest X‐rays using generative adversarial networks | |
Kim et al. | Homogenization of multi-institutional chest x-ray images in various data transformation schemes | |
Przelaskowski et al. | Semantic Segmentation of Abnormal Lung Areas on Chest X-rays to Detect COVID-19 | |
Wan et al. | A hierarchical fine-grained classification approach for COVID-19 severity assessment based on CT images |
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: 12857026 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2013549143 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 112012004826 Country of ref document: DE Ref document number: 1120120048261 Country of ref document: DE |
|
ENP | Entry into the national phase |
Ref document number: 1410473 Country of ref document: GB Kind code of ref document: A Free format text: PCT FILING DATE = 20120920 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1410473.1 Country of ref document: GB |
|
WWE | Wipo information: entry into national phase |
Ref document number: 14365337 Country of ref document: US |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 12857026 Country of ref document: EP Kind code of ref document: A1 |