CN102306173A - Image similarity comparison method - Google Patents
Image similarity comparison method Download PDFInfo
- Publication number
- CN102306173A CN102306173A CN201110246403A CN201110246403A CN102306173A CN 102306173 A CN102306173 A CN 102306173A CN 201110246403 A CN201110246403 A CN 201110246403A CN 201110246403 A CN201110246403 A CN 201110246403A CN 102306173 A CN102306173 A CN 102306173A
- Authority
- CN
- China
- Prior art keywords
- gravity
- irregular
- center
- frontier point
- target
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses an image similarity comparison method, and belongs to the field of image retrieval. In the method, a color image is divided into a plurality of irregular graphs, and image similarity comparison is realized by comparing the similarity of the irregular graphs. The method is simple, and easy to implement, can be used for judging whether two images are similar or not, and is applied to the aspect of image retrieval, matching accuracy can be enhanced, and image retrieval efficiency is improved.
Description
Technical field
The present invention relates to a kind of image similarity comparative approach, especially a kind of through comparing the method that irregular figure realizes that image similarity compares.
Background technology
Late nineteen nineties; Be accompanied by the rise of the universal and multimedia application of the growth of the network bandwidth, extensive storage medium; Add picture and on the ability of expressing the meaning, have inborn advantage than literal, picture is applied to the expression of content and the carrying of information by increasing.
In order to search the required image of user the images millions of from network fast and accurately, matching algorithm is most important.The traditional image matching algorithm exists a lot of not enough, such as color histogram common factor method, only considers certain color proportion in the image, and has ignored the space distribution information of shades of colour in image etc.Utilization common effectiveness of retrieval of traditional image matching algorithm and accuracy are all lower in the example image retrieval.
Summary of the invention
The purpose of this invention is to provide a kind of simple more image similarity comparative approach, it is applied to can improve effectiveness of retrieval and accuracy in the example image retrieval.
To achieve these goals, the invention provides a kind of image similarity comparative approach, may further comprise the steps:
S1, first pixel with target image A, the contrast images B upper left corner is that initial point is set up coordinate system respectively, and setting horizontal ordinate X axis right side be positive dirction, and it is positive dirction that ordinate Y axially descends;
S2, with said target image A according to figure in contained color be divided into a plurality of target irregular figures; Said contrast images B is divided into a plurality of contrast irregular figures according to contained color among the figure; Comparison object irregular figure and contrast irregular figure, carry out following steps:
(01) writes down the center of gravity A of said target irregular figure respectively
0Coordinate be (Ax
0, Ay
0), the center of gravity B of said contrast irregular figure
0Coordinate be (Bx
0, By
0), and with two center of gravity coordinate (Ax
0, Ay
0), (Bx
0, By
0) be converted into relative coordinate (AX respectively
0, AY
0), (BX
0, BY
0), AX wherein
0=Ax
0/ L
A, BX
0=Bx
0/ L
B, AY
0=Ay
0/ W
A, BY
0=By
0/ W
B, L
AThe length of expression target image A, W
AThe width of expression target image A; L
BThe length of expression target image B, W
BThe width of expression target image B;
(02) relative coordinate (AX of center of gravity in more said target irregular figure and the said contrast irregular figure
0, AY
0), (BX
0, BY
0), when the difference of two relative coordinates≤first threshold values, represent that both are close, then continue to carry out, otherwise represent said target irregular figure and said contrast irregular figure dissmilarity;
(03) coordinate of frontier point Ai is (Ax in the said target irregular figure of setting
i, Ay
i), the coordinate of frontier point Bj is (Bx in the said relatively irregular figure
j, By
j), judge that said frontier point Ai is with respect to center of gravity A
0Direction whether with said frontier point Bj with respect to center of gravity B
0Direction identical, if identical then continue to carry out following steps, otherwise this step is carried out in circulation, 1≤i≤n wherein, 1≤j≤m, n, m are natural number;
(04) calculates said frontier point Ai to center of gravity A according to the Euclidean distance formula
0Distance be di=
, and said frontier point Bj is to center of gravity B
0Distance be Dj=
, and ask for said frontier point Ai to center of gravity A
0Apart from di and frontier point Bj to center of gravity B
0The ratio r i=di/Dj of distance B j, 1≤i≤n wherein, n is a natural number;
(05) repeating step (03)~(04), the processing of all frontier points in accomplishing said target irregular figure and said comparison irregular figure obtains the sequence Ri of all distance ratio, 1≤i≤n wherein, n is a natural number;
(06) finds out maximum ratio max (Ri) and minimum ratio min (Ri) among the sequence Ri; If maximum ratio max (Ri) deducts 1 less than second threshold values with the ratio of minimum ratio min (Ri); Promptly | < second threshold values representes that then said target irregular figure is similar with said contrast irregular figure to max (Ri)/>min (Ri)-1|;
(07) when all target irregular figures correspondences of dividing and contrast irregular figure all similar, confirms that said target image A is similar with said contrast images B.
Judge in the said step (03) that said frontier point Ai is with respect to center of gravity A
0Direction whether with said frontier point Bj with respect to center of gravity B
0Direction identical, carry out according to following steps:
(030) calculates said frontier point Ai and center of gravity A
0Horizontal ordinate difference, i.e. Δ x=Ax
i-Ax
0, Δ y=Ay
i-Ay
0, confirm the quadrature N of said frontier point Ai
AiCalculate said frontier point Bj and center of gravity B
0Horizontal ordinate difference, i.e. Δ x '=Bx
j-Bx
0, Δ y '=By
j-By
0, confirm the quadrature N of said frontier point Bj
Bj1≤i≤n wherein, 1≤j≤m, n, m are natural number;
(031) calculates said frontier point Ai with respect to center of gravity A
0Tangent value T
Ai=tan θ=| Δ y/ Δ x| calculates said frontier point Bj with respect to center of gravity B
0Tangent value T
Bj=tan θ=| Δ y '/Δ x ' |;
(032) more said tangent value T
Ai, T
BjAnd quadrature N
Ai, N
BjIf, the difference of two tangent values≤the 3rd threshold values, and two quadratures are equal, represent that then said frontier point Ai is with respect to center of gravity A
0Direction and said frontier point Bj with respect to center of gravity B
0Direction identical, continue to carry out, otherwise return execution in step (030).
Being divided into of four different quadratures in the said step (030):
As Δ x>0, Δ y>0 o'clock N
Ai=4, as Δ x<0, Δ y>0 o'clock N
Ai=3, as Δ x<0, Δ y<0 o'clock N
Ai=2, as Δ x>0, Δ y<0 o'clock N
Ai=1;
As Δ x ’>0, Δ y ’>0 o'clock N
Bj=4, as Δ x '<0, Δ y ’>0 o'clock N
Bj=3, as Δ x '<0, Δ y '<0 o'clock N
Bj=2, as Δ x ’>0, Δ y '<0 o'clock N
Bj=1.
When dividing irregular figure, adopt region growing algorithm to realize; Judging whether certain pixel can be referred to certain when zone, and basis for estimation is: the color value of this pixel whether with color value difference≤the 4th threshold values of regional original point.
In sum, owing to adopted technique scheme, the invention has the beneficial effects as follows:
This image similarity comparative approach computing is simple, is divided into a plurality of irregular figures to coloured image, through the similarity determination of each irregular figure being realized the similarity determination of coloured image; Be applied to improve effectiveness of retrieval and accuracy in the example image retrieval.
Description of drawings
The present invention will explain through example and with reference to the mode of accompanying drawing, wherein:
Fig. 1 is a process flow diagram of the present invention;
Fig. 2 is a process flow diagram of judging that whether identical frontier point in two images with respect to the direction of center of gravity;
Fig. 3 is the synoptic diagram of the center of gravity of irregular figure among the present invention;
Fig. 4 is the synoptic diagram of the frontier point of irregular figure among the present invention apart from center of gravity.
Embodiment
Disclosed all characteristics in this instructions, or the step in disclosed all methods or the process except mutually exclusive characteristic and/or the step, all can make up by any way.
Disclosed arbitrary characteristic in this instructions (comprising any accessory claim, summary and accompanying drawing) is only if special narration all can be replaced by other equivalences or the alternative features with similar purpose.That is, only if special narration, each characteristic is an example in a series of equivalences or the similar characteristics.
As shown in Figure 1, this image similarity comparative approach is made up of following steps:
S1, first pixel with target image A, the contrast images B upper left corner is that initial point is set up coordinate system respectively, and setting horizontal ordinate X axis right side be positive dirction, and it is positive dirction that ordinate Y axially descends.Generally; The shape of image is square; Therefore can also target image A when setting up coordinate system, first pixel at other three angles (lower left corner, the upper right corner and the lower right corner) is that initial point is set up coordinate system among the contrast images B: with first pixel in the lower left corner is that initial point is when setting up coordinate system; The horizontal ordinate X axis right side is a positive dirction, and ordinate Y axially goes up and is positive dirction; With first pixel in the upper right corner is initial point when setting up coordinate system, and a horizontal ordinate X axis left side is a positive dirction, and ordinate Y axially is positive dirction down; With first pixel in the lower right corner is initial point when setting up coordinate system, and a horizontal ordinate X axis left side is a positive dirction, and ordinate Y axially goes up and is positive dirction.
S2, with said target image A according to figure in contained color be divided into a plurality of target irregular figures; Said contrast images B is divided into a plurality of contrast irregular figures according to contained color among the figure; Comparison object irregular figure and contrast irregular figure, carry out following steps:
(01) writes down the center of gravity A of said target irregular figure respectively
0Coordinate be (Ax
0, Ay
0), the center of gravity B of said contrast irregular figure
0Coordinate be (Bx
0, By
0), and with two center of gravity coordinate (Ax
0, Ay
0), (Bx
0, By
0) be converted into relative coordinate (AX respectively
0, AY
0), (BX
0, BY
0), AX wherein
0=Ax
0/ L
A, BX
0=Bx
0/ L
B, AY
0=Ay
0/ W
A, BY
0=By
0/ W
B, L
AThe length of expression target image A, W
AThe width of expression target image A; L
BThe length of expression target image B, W
BThe width of expression target image B.
(02) comparison object irregular figure and the relative coordinate (AX that contrasts center of gravity in the irregular figure
0, AY
0), (BX
0, BY
0); When the difference of two relative coordinates≤first threshold values, represent that two center of gravity coordinates are close; Then continue to carry out following steps, otherwise expression target irregular figure is dissimilar with the contrast irregular figure, wherein the set basis user's of first threshold values demand and deciding; The degree of accuracy of the bigger then retrieving images of first threshold values is just low more, and the degree of accuracy ground of the more little then retrieving images of opposite first threshold values is just high more.
(03) coordinate of frontier point Ai is (Ax in the target setting irregular figure
i, Ay
i), relatively the coordinate of frontier point Bj is (Bx in the irregular figure
j, By
j), ask for frontier point Ai with respect to center of gravity A
0Direction and frontier point Bj with respect to center of gravity B
0Direction, judge whether both directions identical, if identical then continue to carry out following steps, otherwise this step is carried out in circulation, 1≤i≤n wherein, 1≤j≤m, n, m are natural number;
Judge that above-mentioned frontier point Ai is with respect to center of gravity A
0Direction whether with frontier point Bj with respect to center of gravity B
0The identical method of direction a lot, such as adopting sine value etc. to combine to represent frontier point with respect to the direction of center of gravity etc. with quadrature.The method that the tangent value of employing comparison frontier point combines with quadrature in the present embodiment is judged the direction of frontier point with respect to center of gravity.Specifically carry out according to following steps, as shown in Figure 2:
(030) calculates said frontier point Ai and center of gravity A
0Horizontal ordinate difference, i.e. Δ x=Ax
i-Ax
0, Δ y=Ay
i-Ay
0, confirm the quadrature N of said frontier point Ai
AiCalculate said frontier point Bj and center of gravity B
0Horizontal ordinate difference, i.e. Δ x '=Bx
j-Bx
0, Δ y '=By
j-By
0, confirm the quadrature N of said frontier point Bj
Bj1≤i≤n wherein, 1≤j≤m, n, m are natural number; The division of quadrature can be adopted multiple mode, and four quadratures specifically are divided in the present embodiment: as Δ x>0, Δ y>0 o'clock N
Ai=4, as Δ x<0, Δ y>0 o'clock N
Ai=3, as Δ x<0, Δ y<0 o'clock N
Ai=2, as Δ x>0, Δ y<0 o'clock N
Ai=1; And as Δ x ’>0, Δ y ’>0 o'clock N
Bj=4, as Δ x '<0, Δ y ’>0 o'clock N
Bj=3, as Δ x '<0, Δ y '<0 o'clock N
Bj=2, as Δ x ’>0, Δ y '<0 o'clock N
Bj=1.Certainly, can also give quadrature different values according to actual needs.
(031) as shown in Figure 4, calculate said frontier point Ai with respect to center of gravity A
0Tangent value T
Ai=tan θ=| Δ y/ Δ x| calculates said frontier point Bj with respect to center of gravity B
0Tangent value T
Bj=tan θ=| Δ y '/Δ x ' |;
(032) the tangent value T of more said frontier point Ai, Bj
Ai, T
BjAnd quadrature N
Ai, N
BjIf, the difference of two tangent values≤the 3rd threshold values, and two quadratures are equal, represent that then said frontier point Ai is with respect to center of gravity A
0Direction and said frontier point Bj with respect to center of gravity B
0Direction identical, continue to carry out, otherwise return execution in step (030).
Likewise, the set basis user's of the 3rd threshold values demand and deciding, the degree of accuracy of the bigger then retrieving images of the 3rd threshold values is just low more, and the degree of accuracy ground of the more little then retrieving images of opposite the 3rd threshold values is just high more.
(04) calculates said frontier point Ai to center of gravity A according to the Euclidean distance formula
0Distance be di=
, and said frontier point Bj is to center of gravity B
0Distance be Dj=
, and ask for said frontier point Ai to center of gravity A
0Apart from di and frontier point Bj to center of gravity B
0The ratio r i=di/Dj of distance B j, 1≤i≤n wherein, n is a natural number;
(05) repeating step (03)~(04), the processing of all frontier points in accomplishing target irregular figure and irregular figure relatively obtains the sequence Ri of all distance ratio, 1≤i≤n wherein, n is a natural number;
(06) finds out maximum ratio max (Ri) and minimum ratio min (Ri) among the sequence Ri; If maximum ratio max (Ri) deducts 1 less than second threshold values with the ratio of minimum ratio min (Ri); Promptly | < second threshold values representes that then said target irregular figure is similar with said contrast irregular figure to max (Ri)/>min (Ri)-1|;
(07) when all target irregular figures correspondences of dividing and contrast irregular figure all similar, confirms that said target image A is similar with said contrast images B.
When dividing irregular figure, adopt region growing algorithm to realize in the present embodiment; When judging whether certain pixel can be referred to certain zone; Basis for estimation is: the color value of this pixel whether with color value difference≤the 4th threshold values of regional original point; The set basis user's of the 4th threshold values demand and deciding wherein; The degree of accuracy of the bigger then retrieving images of the 4th threshold values is just low more, and the degree of accuracy ground of the more little then retrieving images of opposite the 4th threshold values is just high more.
The present invention is not limited to aforesaid embodiment.The present invention expands to any new feature or any new combination that discloses in this manual, and the arbitrary new method that discloses or step or any new combination of process.
Claims (4)
1. image similarity comparative approach is characterized in that may further comprise the steps:
S1, first pixel with target image A, the contrast images B upper left corner is that initial point is set up coordinate system respectively, and setting horizontal ordinate X axis right side be positive dirction, and it is positive dirction that ordinate Y axially descends;
S2, with said target image A according to figure in contained color be divided into a plurality of target irregular figures; Said contrast images B is divided into a plurality of contrast irregular figures according to contained color among the figure; Comparison object irregular figure and contrast irregular figure, carry out following steps:
(01) writes down the center of gravity A of said target irregular figure respectively
0Coordinate be (Ax
0, Ay
0), the center of gravity B of said contrast irregular figure
0Coordinate be (Bx
0, By
0), and with two center of gravity coordinate (Ax
0, Ay
0), (Bx
0, By
0) be converted into relative coordinate (AX respectively
0, AY
0), (BX
0, BY
0), AX wherein
0=Ax
0/ L
A, BX
0=Bx
0/ L
B, AY
0=Ay
0/ W
A, BY
0=By
0/ W
B, L
AThe length of expression target image A, W
AThe width of expression target image A; L
BThe length of expression target image B, W
BThe width of expression target image B;
(02) relative coordinate (AX of center of gravity in more said target irregular figure and the said contrast irregular figure
0, AY
0), (BX
0, BY
0), when the difference of two relative coordinates≤first threshold values, represent that both are close, then continue to carry out, otherwise represent said target irregular figure and said contrast irregular figure dissmilarity;
(03) coordinate of frontier point Ai is (Ax in the said target irregular figure of setting
i, Ay
i), the coordinate of frontier point Bj is (Bx in the said relatively irregular figure
j, By
j), judge that said frontier point Ai is with respect to center of gravity A
0Direction whether with said frontier point Bj with respect to center of gravity B
0Direction identical, if identical then continue to carry out following steps, otherwise this step is carried out in circulation, 1≤i≤n wherein, 1≤j≤m, n, m are natural number;
(04) calculates said frontier point Ai to center of gravity A according to the Euclidean distance formula
0Distance be di=
, and said frontier point Bj is to center of gravity B
0Distance be Dj=
, and ask for said frontier point Ai to center of gravity A
0Apart from di and frontier point Bj to center of gravity B
0The ratio r i=di/Dj of distance B j, 1≤i≤n wherein, n is a natural number;
(05) repeating step (03)~(04), the processing of all frontier points in accomplishing said target irregular figure and said comparison irregular figure obtains the sequence Ri of all distance ratio, 1≤i≤n wherein, n is a natural number;
(06) finds out maximum ratio max (Ri) and minimum ratio min (Ri) among the sequence Ri; If maximum ratio max (Ri) deducts 1 less than second threshold values with the ratio of minimum ratio min (Ri); Promptly | < second threshold values representes that then said target irregular figure is similar with said contrast irregular figure to max (Ri)/>min (Ri)-1|;
(07) when all target irregular figures correspondences of dividing and contrast irregular figure all similar, confirms that said target image A is similar with said contrast images B.
2. image similarity comparative approach according to claim 1 is characterized in that: judge in the said step (03) that said frontier point Ai is with respect to center of gravity A
0Direction whether with said frontier point Bj with respect to center of gravity B
0Direction identical, carry out according to following steps:
(030) calculates said frontier point Ai and center of gravity A
0Horizontal ordinate difference, i.e. Δ x=Ax
i-Ax
0, Δ y=Ay
i-Ay
0, confirm the quadrature N of said frontier point Ai
AiCalculate said frontier point Bj and center of gravity B
0Horizontal ordinate difference, i.e. Δ x '=Bx
j-Bx
0, Δ y '=By
j-By
0, confirm the quadrature N of said frontier point Bj
Bj1≤i≤n wherein, 1≤j≤m, n, m are natural number;
(031) calculates said frontier point Ai with respect to center of gravity A
0Tangent value T
Ai=tan θ=| Δ y/ Δ x| calculates said frontier point Bj with respect to center of gravity B
0Tangent value T
Bj=tan θ=| Δ y '/Δ x ' |;
(032) more said tangent value T
Ai, T
BjAnd quadrature N
Ai, N
BjIf, the difference of two tangent values≤the 3rd threshold values, and two quadratures are equal, represent that then said frontier point Ai is with respect to center of gravity A
0Direction and said frontier point Bj with respect to center of gravity B
0Direction identical, continue to carry out, otherwise return execution in step (030).
3. image similarity comparative approach according to claim 2 is characterized in that: being divided into of four different quadratures in the said step (030):
As Δ x>0, Δ y>0 o'clock N
Ai=4, as Δ x<0, Δ y>0 o'clock N
Ai=3, as Δ x<0, Δ y<0 o'clock N
Ai=2, as Δ x>0, Δ y<0 o'clock N
Ai=1;
As Δ x ’>0, Δ y ’>0 o'clock N
Bj=4, as Δ x '<0, Δ y ’>0 o'clock N
Bj=3, as Δ x '<0, Δ y '<0 o'clock N
Bj=2, as Δ x ’>0, Δ y '<0 o'clock N
Bj=1.
4. image similarity comparative approach according to claim 1 is characterized in that: when dividing irregular figure, adopt region growing algorithm to realize; Judging whether certain pixel can be referred to certain when zone, and basis for estimation is: the color value of this pixel whether with color value difference≤the 4th threshold values of regional original point.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2011102464030A CN102306173B (en) | 2011-08-25 | 2011-08-25 | Image similarity comparison method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2011102464030A CN102306173B (en) | 2011-08-25 | 2011-08-25 | Image similarity comparison method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN102306173A true CN102306173A (en) | 2012-01-04 |
CN102306173B CN102306173B (en) | 2012-11-14 |
Family
ID=45380035
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2011102464030A Expired - Fee Related CN102306173B (en) | 2011-08-25 | 2011-08-25 | Image similarity comparison method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102306173B (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103929624A (en) * | 2014-04-29 | 2014-07-16 | 金三立视频科技(深圳)有限公司 | Box camera and ball camera cooperative tracking and monitoring method |
CN104794139A (en) * | 2014-01-22 | 2015-07-22 | 腾讯科技(北京)有限公司 | Information retrieval method, device and system |
CN105956198A (en) * | 2016-06-20 | 2016-09-21 | 东北大学 | Nidus position and content-based mammary image retrieval system and method |
CN107844803A (en) * | 2017-10-30 | 2018-03-27 | ***股份有限公司 | The method and apparatus that a kind of picture compares |
CN109344868A (en) * | 2018-08-28 | 2019-02-15 | 广东奥普特科技股份有限公司 | A kind of universal method for distinguishing axisymmetric inhomogeneity object each other |
CN109558875A (en) * | 2018-11-14 | 2019-04-02 | 广州同略信息科技有限公司 | Method, apparatus, terminal and storage medium based on image automatic identification |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0898245A1 (en) * | 1997-08-05 | 1999-02-24 | Canon Kabushiki Kaisha | Image processing apparatus |
CN101996325A (en) * | 2010-09-08 | 2011-03-30 | 北京航空航天大学 | Improved method for extracting characteristic point from image |
-
2011
- 2011-08-25 CN CN2011102464030A patent/CN102306173B/en not_active Expired - Fee Related
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0898245A1 (en) * | 1997-08-05 | 1999-02-24 | Canon Kabushiki Kaisha | Image processing apparatus |
CN101996325A (en) * | 2010-09-08 | 2011-03-30 | 北京航空航天大学 | Improved method for extracting characteristic point from image |
Non-Patent Citations (4)
Title |
---|
JUN TANG: "A color image segmentation algorithm based on region growing", 《COMPUTER ENGINEERING AND TECHNOLOGY(ICCET),2010 2ND INTERNATIONAL CONFERENCE》, 18 April 2010 (2010-04-18), pages 634 - 637 * |
乔玲玲: "图像分割算法研究及实现", 《中国优秀硕士学位论文全文数据库信息科技辑》, 15 September 2009 (2009-09-15), pages 1 - 53 * |
党建武等: "基于子区域相似度的医学图像分割算法", 《计算机应用》, vol. 30, no. 9, 30 September 2010 (2010-09-30), pages 2458 - 2460 * |
黄长专等: "图像分割方法研究", 《计算机技术与发展》, vol. 19, no. 6, 30 June 2009 (2009-06-30), pages 76 - 79 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104794139A (en) * | 2014-01-22 | 2015-07-22 | 腾讯科技(北京)有限公司 | Information retrieval method, device and system |
CN104794139B (en) * | 2014-01-22 | 2019-09-20 | 腾讯科技(北京)有限公司 | Information retrieval method, apparatus and system |
CN103929624A (en) * | 2014-04-29 | 2014-07-16 | 金三立视频科技(深圳)有限公司 | Box camera and ball camera cooperative tracking and monitoring method |
CN103929624B (en) * | 2014-04-29 | 2017-07-04 | 深圳金三立视频科技股份有限公司 | Gunlock, ball machine collaboration tracing and monitoring method |
CN105956198A (en) * | 2016-06-20 | 2016-09-21 | 东北大学 | Nidus position and content-based mammary image retrieval system and method |
CN105956198B (en) * | 2016-06-20 | 2019-04-26 | 东北大学 | A kind of galactophore image searching system and method based on lesions position and content |
CN107844803A (en) * | 2017-10-30 | 2018-03-27 | ***股份有限公司 | The method and apparatus that a kind of picture compares |
CN107844803B (en) * | 2017-10-30 | 2021-12-28 | ***股份有限公司 | Picture comparison method and device |
CN109344868A (en) * | 2018-08-28 | 2019-02-15 | 广东奥普特科技股份有限公司 | A kind of universal method for distinguishing axisymmetric inhomogeneity object each other |
CN109344868B (en) * | 2018-08-28 | 2021-11-16 | 广东奥普特科技股份有限公司 | General method for distinguishing different types of objects which are mutually axisymmetric |
CN109558875A (en) * | 2018-11-14 | 2019-04-02 | 广州同略信息科技有限公司 | Method, apparatus, terminal and storage medium based on image automatic identification |
Also Published As
Publication number | Publication date |
---|---|
CN102306173B (en) | 2012-11-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102306173B (en) | Image similarity comparison method | |
JP2020119550A (en) | Graphical reference marker identification suitable for augmented reality, virtual reality, and robotics | |
JP5463415B2 (en) | Method and system for quasi-duplicate image retrieval | |
US8873835B2 (en) | Methods and apparatus for correcting disparity maps using statistical analysis on local neighborhoods | |
Nodehi et al. | Intelligent fuzzy approach for fast fractal image compression | |
CN105513105B (en) | Image background weakening method based on notable figure | |
CN102509099A (en) | Detection method for image salient region | |
CN103768792A (en) | Partitioning and recognition method applied to wide-range video space object positioning system | |
CN103914819B (en) | A kind of based on the infrared image joining method improving RANSAC | |
CN110163894B (en) | Sub-pixel level target tracking method based on feature matching | |
CN104331890A (en) | Method and system for estimating global disparity | |
Yuan et al. | Superpixels with content-adaptive criteria | |
CN104077561B (en) | Fingerprint automatic comparison method | |
CN103336963A (en) | Method and device for image feature extraction | |
CN103116883A (en) | Normalized cross correlation (NCC) registration method of self-adaptation threshold | |
CN111105490B (en) | Rapid normal vector orientation method for scattered point clouds | |
CN104994368B (en) | Non-key frame sort method in 2D 3D Video Quality Metrics | |
CN106408029A (en) | Image texture classification method based on structural difference histogram | |
Huang et al. | Unstructured lane identification based on hough transform and improved region growing | |
CN103679676A (en) | Quick unordered image stitching method based on multi-level word bag clustering | |
Zhi-Yong et al. | The license plate image binarization based on otsu algorithm and MATLAB realize | |
Wang et al. | Road detection via superpixels and interactive image segmentation | |
CN110852352A (en) | Data enhancement method for training deep neural network model for target detection | |
Walters et al. | ChromaTag-A Colored Fiducial Marker | |
Lihua et al. | 3D point cloud registration for apple tree based on Kinect camera |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
C17 | Cessation of patent right | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20121114 Termination date: 20130825 |