CN105574837A - Image similarity matching method and device - Google Patents
Image similarity matching method and device Download PDFInfo
- Publication number
- CN105574837A CN105574837A CN201410539457.XA CN201410539457A CN105574837A CN 105574837 A CN105574837 A CN 105574837A CN 201410539457 A CN201410539457 A CN 201410539457A CN 105574837 A CN105574837 A CN 105574837A
- Authority
- CN
- China
- Prior art keywords
- image
- channel
- similarity
- matched
- channel image
- 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.)
- Pending
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention is suitable of the image processing field, and provides an image similarity matching method and device. The method comprises the following steps: obtaining the main distribution characteristics of a color area in a colorful image to be matched; reducing an R channel image, a G channel image and a B channel image in the colorful image to be matched; coding the reduced R channel image, G channel image and B channel image to obtain corresponding coding values; comparing the main distribution characteristics of the colorful areas in the colorful image to be matched with the main distribution characteristics of the colorful areas in a stored image to obtain a first similarity; comparing the coding values of the reduced R channel image, G channel image and B channel image with the coding value of the stored image to obtain a second similarity; and according to the first similarity and the second similarity, judging whether the colorful image to be matched is matched with the stored image or not. The embodiment of the invention can improve matching accuracy.
Description
Technical field
The invention belongs to image processing field, particularly relate to a kind of image similarity match method and device.
Background technology
Images match, mainly by corresponding relation, the similarity of the texture of analysis two images and gray scale etc., judges that whether these two images are identical.
At present, in the process of carrying out image similarity match, all often undertaken getting description to picture material structure based on pixel operation by the gray level image corresponding to coloured image, as scale invariant feature conversion (Scale-invariantfeaturetransform, SIFT) algorithm, local binary patterns (LocalBinaryPatterns, LBP) algorithm, gray level co-occurrence matrixes (GrayLevelConcurrenceMatrix) algorithm etc., but in method recited above, there are himself deficiency and defect, although as fine in SIFT robustness, anti-deformation, anti-rotation performance is strong, but calculated amount is very large, in the characteristic matching of a lot of high-definition image, real-time is very poor, LBP algorithm and gray level co-occurrence matrixes are all based on gray level image, and the colouring information of other passage is eliminated due to gray level image, therefore a lot of foundation has been lacked by the similarity of LBP algorithm and gray level co-occurrence matrixes method comparison image, easily cause False Rate high.
Summary of the invention
Embodiments provide a kind of image similarity match method, be intended to solve the existing method problem that False Rate is too high when matching image.
The embodiment of the present invention is achieved in that a kind of image similarity match method, and described method comprises the steps:
Obtain the main distribution characteristics of color region in coloured image to be matched;
Reduce R channel image in coloured image to be matched, G channel image and channel B image;
Described R channel image, G channel image and channel B image after reducing are encoded, obtains corresponding encoded radio;
The main distribution characteristics of color region in coloured image to be matched is compared with the main distribution characteristics of color region in the image of storage, obtains the first similarity;
Described R channel image, G channel image and the encoded radio of channel B image after reducing are compared with the encoded radio of the image of storage, obtains the second similarity;
Judge whether coloured image to be matched mates with the image of storage according to described first similarity and described second similarity.
Another object of the embodiment of the present invention is to provide a kind of image similarity match device, and described device comprises:
Color Distribution Features acquiring unit, for obtaining the main distribution characteristics of color region in coloured image to be matched;
Channel image reducing unit, for reducing R channel image in coloured image to be matched, G channel image and channel B image;
Channel image coding unit, for encoding to described R channel image, G channel image and the channel B image after reducing, obtains corresponding encoded radio;
First similarity determining unit, for the main distribution characteristics of color region in coloured image to be matched being compared with the main distribution characteristics of color region in the image of storage, obtains the first similarity;
Second similarity determining unit, for being compared with the encoded radio of the image of storage by described R channel image, G channel image and the encoded radio of channel B image after reducing, obtains the second similarity;
According to described first similarity and described second similarity, image matching unit, for judging whether coloured image to be matched mates with the image of storage.
In embodiments of the present invention, owing to remaining the colouring information of three passages in matching process, add match information, therefore, it is possible to improve the accuracy rate of coupling.And, due to the histogram that the first similarity is based on piecemeal color, the second similarity is image content-based structure, therefore, the image similarity match methods combining that the embodiment of the present invention provides piecemeal color histogram similarity and picture material structural similarity, thus make coupling more accurate.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of a kind of image similarity match method that first embodiment of the invention provides;
Fig. 2 is the schematic diagram extracting R channel image, G channel image, channel B image from coloured image respectively that first embodiment of the invention provides;
Fig. 3 is the histogram of the R channel image shown in Fig. 2, G channel image and the channel B image that first embodiment of the invention provides;
Fig. 4 is the histogram obtained after the reduced graph 3 that provides of first embodiment of the invention;
Fig. 5 be first embodiment of the invention provide reduce Fig. 2 after the schematic diagram that obtains;
Fig. 6 is the schematic diagram of encoding to the image of 15*15 pixel size that first embodiment of the invention provides;
Fig. 7 is the structural drawing of a kind of image similarity match device that second embodiment of the invention provides.
Embodiment
In order to make object of the present invention, technical scheme and advantage clearly understand, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.
In the embodiment of the present invention, obtain the main distribution characteristics of color region in coloured image to be matched, and to red in coloured image to be matched (R) channel image, green (G) channel image and indigo plant (B) channel image are encoded respectively, obtain corresponding encoded radio, again respectively by the main distribution characteristics of color region in coloured image to be matched, the main distribution characteristics of color region in the image of encoded radio and storage, encoded radio compares, determine the first similarity and the second similarity, finally judge whether coloured image to be matched mates with the image of storage according to the first similarity determined and the second similarity.
In order to technical solutions according to the invention are described, be described below by specific embodiment.
embodiment one:
Fig. 1 shows the process flow diagram of a kind of image similarity match method that first embodiment of the invention provides, and details are as follows:
Step S11, obtains the main distribution characteristics of color region in coloured image to be matched.
Wherein, in described acquisition coloured image to be matched, the step of the main distribution characteristics of color region specifically comprises:
A1, from coloured image to be matched, extract R channel image, G channel image, channel B image respectively.As shown in Figure 2, R channel image, G channel image, channel B image can be extracted respectively in a coloured image to be matched.
A2, respectively R channel image, G channel image and channel B image are divided into equal-sized piece.Particularly, should ensure that the pixel value size of the block of each division is more than or equal to 20*20, i.e. the block pixel value in the horizontal direction of each division is more than or equal to 20, and pixel value is in the vertical direction more than or equal to 20.Ensure that the size of the block divided is because only have a certain amount of image pixel fully could describe out the color distribution situation of one piece of picture material.
A3, determine divide the histogram information of all pieces.Particularly, because image generally includes 256 pixel values, therefore, choose a numerical value of N (N is less than or equal to 255), by the N chosen, 0 ~ 255 these 256 pixels are divided into N part.Suppose 256/N=Path_num, if Path_num is not integer, then choose the max-int being less than Path_num; If integer, then directly choose this integer of Path_num.When 256 pixel values being divided into N number of interval, institute's by stages rule is as follows: work as i=1, and pixel value arrives a conduct interval of (Path_num*i) 0; As 2≤i≤N-1, pixel value is interval as N-1 at Path_num* (i-1)+1 to Path_num*i; Work as i=N, pixel value (Path_num* (i-1)+1) to 255 as an interval.Such as, suppose N=10, so there is 256/10=25.6, get the max-int 25 being less than 25.6, i.e. Patch_num=25, so for pixel value in image 0 ~ 25 is a statistical pixel interval, pixel value be 26 ~ 50 be between a Statistical Area, by that analogy until 201 ~ 225,226 ~ 255.The histogram of the R channel image shown in Fig. 3, G channel image and channel B image can be obtained according to the method, refer to Fig. 3.
A4, from the histogram informations of described all pieces, extract the main distribution characteristics of color region in R channel image, G channel image and channel B image respectively, and synthesize the main distribution characteristics of color region in coloured image to be matched.Particularly, this steps A 4 comprises: A41, sort to the size of the statistical value of each pixel range in the histogram information in R channel image, G channel image and channel B image respectively; The statistical value of the maximum pixel range of specified quantity in A42, the histogram information retaining R channel image, the histogram information of G channel image and the histogram information of channel B image, and retain the left side of statistical value of pixel range and the statistical value of the pixel range on the right that are retained in the histogram information of R channel image, the histogram information of G channel image and the histogram information of channel B image, using the main distribution characteristics as color region in R channel image, G channel image and channel B image; In this step, suppose 256 pixel values to be divided into N number of part, then retain maximum N*50% interval, and retain the left side in these intervals and the interval on the right.Suppose N=10, then select maximum 10*50%=5 interval, and retain the interval on the interval left side of these maximum 5 and the right, continue for Fig. 3, then after the histogram of a block of pixels removes partial section, the histogram of the interval composition of reservation as shown in Figure 4.As can be seen from the Color Channel of three in Fig. 4, the distribution situation of the color region of a block of pixels, interval value is larger, illustrates that, in this block of pixels, the pixel value of this scope accounts for major part.Finally, to each piecemeal of tri-passages of R, G, B in Fig. 3 according to the distribution sequence of block of pixels, from left to right, the main distribution characteristics of the color region of three color channel image is determined from top to bottom.A43, by the main distribution characteristics of color region in R channel image, G channel image and channel B image, synthesize the main distribution characteristics of color region in coloured image to be matched in order.In this step, according to the order of the main distribution characteristics of color region in R, G, B tri-Color Channels, synthesize the main distribution characteristics of color region in coloured image to be matched.
Step S12, reduces R channel image in coloured image to be matched, G channel image and channel B image.
In this step, be the subject information of only remaining image outline by R channel image, G channel image and channel B image down in coloured image to be matched, substantially there is no the image of detailed information.Such as, can by image down in the scope of 20*20 pixel size, or, by image down to 15*15 pixel size.Continue for Fig. 2, then, after the reduction process by large scale, obtain image as shown in Figure 5, this Fig. 5 only remains basic image outline, does not substantially have the description of details.This step, by reducing red R channel image in coloured image to be matched, green G channel image and blue channel B image, decreases follow-up calculated amount, thus improves the speed of images match.
Step S13, encodes to described R channel image, G channel image and channel B image after reducing, obtains corresponding encoded radio.
Wherein, describedly to encode to described R channel image, G channel image and the channel B image after reducing, the step obtaining corresponding encoded radio specifically comprises:
Following coding is all performed to described R channel image, G channel image and channel B image after reducing:
B1, from the block of pixels in the upper left corner of channel image, from left to right, from top to bottom, two more adjacent in order pixel values.
B2, carry out corresponding mark according to comparative result, described in be designated encoded radio.
In order to be illustrated more clearly in the step of B1 and B2, being described with an object lesson below: refer to Fig. 6, supposing that by R channel image, G channel image and channel B image down be 15*15 pixel size.From the upper left corner of the image after this reduces, the position mark of first block of pixels is (1,1) position mark of the block of pixels that, in a lateral direction, moves right is (1,2), the at this time pixel value size of compared pixels block (1,1) and block of pixels (1,2), if block of pixels (1,1) pixel value be more than or equal to the pixel value of block of pixels (1,2), then be labeled as 1 (namely encoded radio is 1), if the pixel value of block of pixels (1,1) is less than block of pixels (1,2) pixel value, be then labeled as 0 (namely encoded radio is 0).Complete one-row pixels block relatively after, last block of pixels (1 of the first row, 15) with first block of pixels (2 of the second row, 1) continue in the manner described above to compare, repeat above-mentioned steps, until complete the comparison of all block of pixels, finally obtain the encoded radio of (w_k-1) * (h_k-1) size, wherein, w_k is the number of pixels after image down on cross direction, and h_k is the number of pixels after image down on high direction.
Step S14, compares the main distribution characteristics of color region in coloured image to be matched with the main distribution characteristics of color region in the image of storage, obtains the first similarity.
In this step, the method for histogram intersection can be adopted to judge the similarity of the image of coloured image to be matched and storage in three channel image.
Step S15, compares described R channel image, G channel image and the encoded radio of channel B image after reducing with the encoded radio of the image of storage, obtains the second similarity.
In this step, the encoded radio of the encoded radio of R channel image, the encoded radio of G channel image and the channel B image after reducing is compared with the encoded radio of the encoded radio of R channel image of the image stored, the encoded radio of G channel image and channel B image respectively, when specifically comparing, be whether the encoded radio compared in same position is identical, determine the second similarity finally by method as described below: the ratio of same position same-code value and total code length value.
According to described first similarity and described second similarity, step S16, judges whether coloured image to be matched mates with the image of storage.
Wherein, describedly judge that the step whether coloured image to be matched and the image of storage mate specifically comprises according to described first similarity and described second similarity:
C1, determine the mean value of described first similarity and described second similarity.
Whether the mean value that C2, judgement obtain is greater than the threshold value preset, and if so, judges the images match of coloured image to be matched and storage, if not, judges that coloured image to be matched does not mate with the image of storage.
In first embodiment of the invention, obtain the main distribution characteristics of color region in coloured image to be matched, and to R channel image red in coloured image to be matched, green G channel image and blue channel B image are encoded respectively, obtain corresponding encoded radio, again respectively by the main distribution characteristics of color region in coloured image to be matched, the main distribution characteristics of color region in the image of encoded radio and storage, encoded radio compares, determine the first similarity and the second similarity, finally judge whether coloured image to be matched mates with the image of storage according to the first similarity determined and the second similarity.Owing to remaining the colouring information of three passages in matching process, add match information amount, therefore, it is possible to improve the accuracy rate of coupling.And, due to the histogram that the first similarity is based on piecemeal color, the second similarity is image content-based structure, therefore, the image similarity match methods combining that the embodiment of the present invention provides piecemeal color histogram similarity and picture material structural similarity, thus make coupling more accurate.
embodiment two:
Fig. 7 shows the structural drawing of a kind of image similarity match device that second embodiment of the invention provides, and for convenience of explanation, illustrate only the part relevant to the embodiment of the present invention.
Described image similarity match device comprises: Color Distribution Features acquiring unit 71, channel image reducing unit 72, channel image coding unit 73, first similarity determining unit 74, second similarity determining unit 75, image matching unit 76.Wherein:
Color Distribution Features acquiring unit 71, for obtaining the main distribution characteristics of color region in coloured image to be matched.
Wherein, described Color Distribution Features acquiring unit 71 comprises: channel image extraction module, block of pixels divide the Color Distribution Features synthesis module of module, histogram information determination module, image.
This channel image extraction module, for extracting R channel image, G channel image, channel B image respectively from coloured image to be matched.
This block of pixels divides module, for respectively R channel image, G channel image and channel B image being divided into equal-sized piece.Particularly, should ensure that the pixel value size of the block of each division is more than or equal to 20*20
This histogram information determination module, for determining the histogram information of all pieces divided.Particularly, choose a numerical value of N (N is less than or equal to 255), by the N chosen, 0 ~ 255 these 256 pixels are divided into N part.Suppose 256/N=Path_num, if Path_num is not integer, then choose the max-int being less than Path_num; If integer, then directly choose this integer of Path_num.
The Color Distribution Features synthesis module of this image, for extracting the main distribution characteristics of color region in R channel image, G channel image and channel B image from the histogram information of described all pieces respectively, and synthesize the main distribution characteristics of color region in coloured image to be matched.Particularly, the Color Distribution Features synthesis module of described image comprises: interval statistics value order module, for sorting to the size of the statistical value of each pixel range in the histogram information in R channel image, G channel image and channel B image respectively; Interval statistics value rejects module, for retain the histogram information of R channel image, the histogram information of G channel image and channel B image histogram information in the statistical value of maximum pixel range of specified quantity, and retain the left side of statistical value of pixel range and the statistical value of the pixel range on the right that are retained in the histogram information of R channel image, the histogram information of G channel image and the histogram information of channel B image, using the main distribution characteristics as color region in R channel image, G channel image and channel B image; Such as, suppose 256 pixel values to be divided into N number of part, then retain maximum N*50% interval, and retain the left side in these intervals and the interval on the right.Three-channel Color Distribution Features synthesis module, for the main distribution characteristics by color region in R channel image, G channel image and channel B image, synthesizes the main distribution characteristics of color region in coloured image to be matched in order.Here order is synthesized according to the order of the main distribution characteristics of color region in R, G, B tri-Color Channels.
Channel image reducing unit 72, for reducing red R channel image in coloured image to be matched, green G channel image and blue channel B image.
Wherein, the scope reduced is: be the subject information of only remaining image outline by R channel image, G channel image and channel B image down in coloured image to be matched, does not substantially have the image of detailed information.
Channel image coding unit 73, for encoding to described R channel image, G channel image and the channel B image after reducing, obtains corresponding encoded radio.
Wherein, described channel image coding unit comprises: adjacent pixel values comparison module and encoded radio determination module.
Following adjacent pixel values comparison module and encoded radio determination module are all performed to described R channel image, G channel image and channel B image after reducing:
This adjacent pixel values comparison module, the block of pixels for the upper left corner from channel image, from left to right, from top to bottom, two more adjacent in order pixel values.
This encoded radio determination module, for carrying out corresponding mark according to comparative result, described in be designated encoded radio.
First similarity determining unit 74, for the main distribution characteristics of color region in coloured image to be matched being compared with the main distribution characteristics of color region in the image of storage, obtains the first similarity.
Second similarity determining unit 75, for being compared with the encoded radio of the image of storage by described R channel image, G channel image and the encoded radio of channel B image after reducing, obtains the second similarity.
Wherein, when specifically comparing, be whether the encoded radio compared in same position is identical, determine the second similarity finally by following formula: the ratio of same position same-code value and total code length value.
According to described first similarity and described second similarity, image matching unit 76, for judging whether coloured image to be matched mates with the image of storage.
Wherein, described image matching unit 76 comprises:
Similarity mean value determination module, for determining the mean value of described first similarity and described second similarity.
Similarity mean value and threshold value comparison module, for judging whether the mean value obtained is greater than the threshold value preset, and if so, judges the images match of coloured image to be matched and storage, if not, judge that coloured image to be matched does not mate with the image of storage.
In the embodiment of the present invention, owing to remaining the colouring information of three passages in matching process, add coupling and, therefore, it is possible to improve coupling accuracy rate.And, due to the histogram that the first similarity is based on piecemeal color, the second similarity is image content-based structure, therefore, the image similarity match methods combining that the embodiment of the present invention provides piecemeal color histogram similarity and picture material structural similarity, thus make coupling more accurate.
One of ordinary skill in the art will appreciate that, the all or part of step realized in above-described embodiment method is that the hardware that can carry out instruction relevant by program has come, described program can be stored in a computer read/write memory medium, described storage medium, as ROM/RAM, disk, CD etc.
The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, all any amendments done within the spirit and principles in the present invention, equivalent replacement and improvement etc., all should be included within protection scope of the present invention.
Claims (10)
1. an image similarity match method, is characterized in that, described method comprises the steps:
Obtain the main distribution characteristics of color region in coloured image to be matched;
Reduce R channel image in coloured image to be matched, G channel image and channel B image;
Described R channel image, G channel image and channel B image after reducing are encoded, obtains corresponding encoded radio;
The main distribution characteristics of color region in coloured image to be matched is compared with the main distribution characteristics of color region in the image of storage, obtains the first similarity;
Described R channel image, G channel image and the encoded radio of channel B image after reducing are compared with the encoded radio of the image of storage, obtains the second similarity;
Judge whether coloured image to be matched mates with the image of storage according to described first similarity and described second similarity.
2. the method for claim 1, is characterized in that, in described acquisition coloured image to be matched, the step of the main distribution characteristics of color region specifically comprises:
R channel image, G channel image, channel B image is extracted respectively from coloured image to be matched;
Respectively R channel image, G channel image and channel B image are divided into equal-sized piece;
Determine the histogram information of all pieces divided;
From the histogram informations of described all pieces, extract the main distribution characteristics of color region in R channel image, G channel image and channel B image respectively, and synthesize the main distribution characteristics of color region in coloured image to be matched.
3. method as claimed in claim 2, it is characterized in that, the described main distribution characteristics extracting color region in R channel image, G channel image and channel B image from the histogram information of described all pieces respectively, and the step synthesizing the main distribution characteristics of color region in coloured image to be matched specifically comprises: respectively the size of the statistical value of each pixel range in the histogram information in R channel image, G channel image and channel B image is sorted;
Retain the statistical value of the maximum pixel range of specified quantity in the histogram information of R channel image, the histogram information of G channel image and the histogram information of channel B image, and retain the left side of statistical value of pixel range and the statistical value of the pixel range on the right that are retained in the histogram information of R channel image, the histogram information of G channel image and the histogram information of channel B image, using the main distribution characteristics as color region in R channel image, G channel image and channel B image;
By the main distribution characteristics of color region in R channel image, G channel image and channel B image, synthesize the main distribution characteristics of color region in coloured image to be matched in order.
4. the method for claim 1, is characterized in that, describedly encodes to described R channel image, G channel image and the channel B image after reducing, and the step obtaining corresponding encoded radio specifically comprises:
Following coding is all performed to described R channel image, G channel image and channel B image after reducing:
From the block of pixels in the upper left corner of channel image, from left to right, from top to bottom, two more adjacent in order pixel values;
Carry out corresponding mark according to comparative result, described in be designated encoded radio.
5. the method for claim 1, is characterized in that, describedly judges that the step whether coloured image to be matched and the image of storage mate specifically comprises according to described first similarity and described second similarity:
Determine the mean value of described first similarity and described second similarity;
Judge whether the mean value obtained is greater than the threshold value preset, and if so, judges the images match of coloured image to be matched and storage, if not, judges that coloured image to be matched does not mate with the image of storage.
6. an image similarity match device, is characterized in that, described device comprises:
Color Distribution Features acquiring unit, for obtaining the main distribution characteristics of color region in coloured image to be matched;
Channel image reducing unit, for reducing R channel image in coloured image to be matched, G channel image and channel B image;
Channel image coding unit, for encoding to described R channel image, G channel image and the channel B image after reducing, obtains corresponding encoded radio;
First similarity determining unit, for the main distribution characteristics of color region in coloured image to be matched being compared with the main distribution characteristics of color region in the image of storage, obtains the first similarity;
Second similarity determining unit, for being compared with the encoded radio of the image of storage by described R channel image, G channel image and the encoded radio of channel B image after reducing, obtains the second similarity;
According to described first similarity and described second similarity, image matching unit, for judging whether coloured image to be matched mates with the image of storage.
7. device as claimed in claim 6, it is characterized in that, described Color Distribution Features acquiring unit comprises:
Channel image extraction module, for extracting R channel image, G channel image, channel B image respectively from coloured image to be matched;
Block of pixels divides module, for respectively R channel image, G channel image and channel B image being divided into equal-sized piece;
Histogram information determination module, for determining the histogram information of all pieces divided;
The Color Distribution Features synthesis module of image, for extracting the main distribution characteristics of color region in R channel image, G channel image and channel B image from the histogram information of described all pieces respectively, and synthesize the main distribution characteristics of color region in coloured image to be matched.
8. device as claimed in claim 7, it is characterized in that, the Color Distribution Features synthesis module of described image comprises:
Interval statistics value order module, for sorting to the size of the statistical value of each pixel range in the histogram information in R channel image, G channel image and channel B image respectively;
Interval statistics value rejects module, for retain the histogram information of R channel image, the histogram information of G channel image and channel B image histogram information in the statistical value of maximum pixel range of specified quantity, and retain the left side of statistical value of pixel range and the statistical value of the pixel range on the right that are retained in the histogram information of R channel image, the histogram information of G channel image and the histogram information of channel B image, using the main distribution characteristics as color region in R channel image, G channel image and channel B image;
Three-channel Color Distribution Features synthesis module, for the main distribution characteristics by color region in R channel image, G channel image and channel B image, synthesizes the main distribution characteristics of color region in coloured image to be matched in order.
9. device as claimed in claim 6, it is characterized in that, described channel image coding unit comprises:
Described R channel image, G channel image and channel B image after reducing all are performed with lower module:
Adjacent pixel values comparison module, the block of pixels for the upper left corner from channel image, from left to right, from top to bottom, two more adjacent in order pixel values;
Encoded radio determination module, for carrying out corresponding mark according to comparative result, described in be designated encoded radio.
10. device as claimed in claim 6, it is characterized in that, described image matching unit comprises:
Similarity mean value determination module, for determining the mean value of described first similarity and described second similarity;
Similarity mean value and threshold value comparison module, for judging whether the mean value obtained is greater than the threshold value preset, and if so, judges the images match of coloured image to be matched and storage, if not, judge that coloured image to be matched does not mate with the image of storage.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410539457.XA CN105574837A (en) | 2014-10-13 | 2014-10-13 | Image similarity matching method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410539457.XA CN105574837A (en) | 2014-10-13 | 2014-10-13 | Image similarity matching method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105574837A true CN105574837A (en) | 2016-05-11 |
Family
ID=55884932
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410539457.XA Pending CN105574837A (en) | 2014-10-13 | 2014-10-13 | Image similarity matching method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105574837A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109215004A (en) * | 2018-07-26 | 2019-01-15 | 努比亚技术有限公司 | A kind of image composition method, mobile terminal and computer readable storage medium |
| CN110458232A (en) * | 2019-08-13 | 2019-11-15 | 腾讯科技（深圳）有限公司 | A kind of method and apparatus of determining image style similarity |
| CN112184689A (en) * | 2020-10-12 | 2021-01-05 | 罗建华 | Semiconductor device detection method and device, intelligent terminal and storage medium |
| CN113902760B (en) * | 2021-10-19 | 2022-05-17 | 深圳市飘飘宝贝有限公司 | Object edge optimization method, system, device and storage medium in video segmentation |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8023731B2 (en) * | 2006-02-08 | 2011-09-20 | Samsung Electronics Co., Ltd | Apparatus and method for histogram analysis of image and luminance compensation apparatus using the same |
| CN102509118A (en) * | 2011-09-28 | 2012-06-20 | 安科智慧城市技术（中国）有限公司 | Method for monitoring video retrieval |
| CN103440646A (en) * | 2013-08-19 | 2013-12-11 | 成都品果科技有限公司 | Similarity obtaining method for color distribution and texture distribution image retrieval |
-
2014
- 2014-10-13 CN CN201410539457.XA patent/CN105574837A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8023731B2 (en) * | 2006-02-08 | 2011-09-20 | Samsung Electronics Co., Ltd | Apparatus and method for histogram analysis of image and luminance compensation apparatus using the same |
| CN102509118A (en) * | 2011-09-28 | 2012-06-20 | 安科智慧城市技术（中国）有限公司 | Method for monitoring video retrieval |
| CN103440646A (en) * | 2013-08-19 | 2013-12-11 | 成都品果科技有限公司 | Similarity obtaining method for color distribution and texture distribution image retrieval |
Non-Patent Citations (4)
| Title |
|---|
| SUSHILA AGHAV等: "Mitigation of Rotational Constraints in Image Based Plagiarism Detection Using Perceptual Hash", 《INTERNATIONAL JOURNAL OF COMPUTER SCIENCE TRENDS AND TECHNOLOGY》 * |
| 李建生: "图像元数据特征提取及其在检索中的应用", 《中国优秀博硕士学位论文全文数据库(博士)信息科技辑》 * |
| 陈伟平: "图像相似性计算及其GPU加速的若干研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 黄元元等: "一种基于颜色特征的图像检索方法", 《中国图象图形学报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109215004A (en) * | 2018-07-26 | 2019-01-15 | 努比亚技术有限公司 | A kind of image composition method, mobile terminal and computer readable storage medium |
| CN109215004B (en) * | 2018-07-26 | 2021-11-19 | 努比亚技术有限公司 | Image synthesis method, mobile terminal and computer readable storage medium |
| CN110458232A (en) * | 2019-08-13 | 2019-11-15 | 腾讯科技（深圳）有限公司 | A kind of method and apparatus of determining image style similarity |
| CN110458232B (en) * | 2019-08-13 | 2023-05-30 | 腾讯科技（深圳）有限公司 | Method and equipment for determining image style similarity |
| CN112184689A (en) * | 2020-10-12 | 2021-01-05 | 罗建华 | Semiconductor device detection method and device, intelligent terminal and storage medium |
| CN113902760B (en) * | 2021-10-19 | 2022-05-17 | 深圳市飘飘宝贝有限公司 | Object edge optimization method, system, device and storage medium in video segmentation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shahrian et al. | Weighted color and texture sample selection for image matting | |
| US9275300B2 (en) | Method and apparatus for generating image description vector, image detection method and apparatus | |
| CN109344701A (en) | A kind of dynamic gesture identification method based on Kinect | |
| CN105205488B (en) | Word area detection method based on Harris angle points and stroke width | |
| CN113139521B (en) | Pedestrian boundary crossing monitoring method for electric power monitoring | |
| CN104978567B (en) | Vehicle checking method based on scene classification | |
| CN106548160A (en) | A kind of face smile detection method | |
| CN109740572A (en) | A kind of human face in-vivo detection method based on partial color textural characteristics | |
| CN101655983B (en) | Device and method for exacting dominant color | |
| CN105913070B (en) | A kind of multi thread conspicuousness extracting method based on light-field camera | |
| CN109325507B (en) | Image classification method and system combining super-pixel saliency features and HOG features | |
| CN103473551A (en) | Station logo recognition method and system based on SIFT operators | |
| CN102915544A (en) | Video image motion target extracting method based on pattern detection and color segmentation | |
| CN105574837A (en) | Image similarity matching method and device | |
| CN103295013A (en) | Pared area based single-image shadow detection method | |
| CN113724231A (en) | Industrial defect detection method based on semantic segmentation and target detection fusion model | |
| CN106157323A (en) | The insulator division and extracting method that a kind of dynamic division threshold value and block search combine | |
| CN111539330B (en) | Transformer substation digital display instrument identification method based on double-SVM multi-classifier | |
| CN104966348B (en) | A kind of bill images key element integrality detection method and system | |
| CN110472484A (en) | Video key frame extracting method, system and equipment based on multiple view feature | |
| CN107730508A (en) | Color documents images multichannel binary processing method | |
| CN111260645A (en) | Method and system for detecting tampered image based on block classification deep learning | |
| CN104636708A (en) | Partial document image comparison method and system | |
| CN108537771B (en) | MC-SILTP moving target detection method based on HSV | |
| CN106815580A (en) | A kind of license plate locating method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C41 | Transfer of patent application or patent right or utility model | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20160531 Address after: 518000 Guangdong city of Shenzhen province Nanshan District Xili liuxiandong Zhongshan Road No. 1001 TCL Science Park Research Building D4 Building 8 layer B1 unit 802-1. Applicant after: SHENZHEN TCL HIGH-TECH DEVELOPMENT CO., LTD. Address before: 516006 Guangdong province Huizhou Zhongkai hi tech Development Zone No. nineteen District Applicant before: TCL Group Co., Ltd. |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160511 |