CN104112284A - Method and equipment for detecting similarity of images - Google Patents
Method and equipment for detecting similarity of images Download PDFInfo
- Publication number
- CN104112284A CN104112284A CN201310140673.2A CN201310140673A CN104112284A CN 104112284 A CN104112284 A CN 104112284A CN 201310140673 A CN201310140673 A CN 201310140673A CN 104112284 A CN104112284 A CN 104112284A
- Authority
- CN
- China
- Prior art keywords
- measured
- picture
- target photo
- picture block
- block
- 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
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000001514 detection method Methods 0.000 claims abstract description 38
- 238000005520 cutting process Methods 0.000 claims description 74
- 238000013507 mapping Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 description 21
- 238000010586 diagram Methods 0.000 description 11
- 238000003860 storage Methods 0.000 description 9
- 230000008859 change Effects 0.000 description 7
- 238000004590 computer program Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 7
- 238000012360 testing method Methods 0.000 description 7
- 238000012986 modification Methods 0.000 description 6
- 230000004048 modification Effects 0.000 description 6
- 239000000284 extract Substances 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000003556 assay Methods 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000003252 repetitive effect Effects 0.000 description 2
- 241000931705 Cicada Species 0.000 description 1
- 241000283080 Proboscidea <mammal> Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Abstract
The invention discloses a method and a piece of equipment for detecting the similarity of images. The method mainly comprises the following steps: a to-be-detected image and a target image are respectively divided into a plurality of target image blocks; MD5 similarity detection between the images is converted into MD5 similarity detection between the target image blocks; and if the MD5 codes of two image blocks in similarity detection are the same, the contents in the same areas of the target image and the to-be-detected image are completely consistent, and that the target image and the to-be-detected image are similar images can be determined. According to the scheme of the invention, as that two images are similar is determined only when that contents of partial areas of two similar images are completely consistent is determined, the accuracy of detection is high. Meanwhile, as MD5 codes of two images do not need to be the same, but the MD5 codes of the same areas needs to the same, a locally-changed duplicate image can be detected.
Description
Technical field
The application relates to technical field of image processing, relates in particular to a kind of similarity detection method and equipment of picture.
Background technology
Development along with network technology, the content of showing website by issue picture in various websites, for example, shopping website is shown the commodity of sale by the picture of publishing commodity, library's Website shows by issue book cover the books that can borrow, and news website is vividly shown news content by the picture of the event of releasing news.
Picture has played positive role for the information issue of website, still, in website, may have in a large number identical picture, or picture is carried out to the close picture that local modification (such as increasing local word, pattern and watermark etc.) produces afterwards.These reusable identical or close pictures not only can not effectively be shown the information of website issue, also can additionally take the system resource storage space etc. of website (as take) of website, and affects user's experience etc.For example, on shopping website, the commodity of counterweight recurrence cloth are used identical picture, cause the problem of shopping website to the information issuing process additional allocation resource repeating, and concerning consumer, search out be all repetition product, user experiences and can reduce.
For reusable identical or close picture in website is investigated, picture in website and tested picture can be carried out to similarity detection, find out reusable tested picture in website, afterwards, can screen reusable tested picture, remove unnecessary picture.
Conventional picture analogies degree detection technique has following two kinds at present:
The first picture analogies degree detection technique: characteristic similarity detects.
By calculating special characteristics such as histogram, local feature point on two pictures, by machine learning algorithm, the special characteristic on two pictures is carried out to similarity judgement, if similarity reaches threshold value, represent that this two pictures is similar pictures; Otherwise, represent that this two pictures is dissimilar.
Characteristic similarity detection technique can only detect specific feature, because these features may not represent the content of picture accurately, cause two larger pictures of difference all may obtain similar result of determination, this False Rate is higher, does not reach and finds the object that is close to similar pictures; Meanwhile, owing to directly the content in picture being carried out to similarity judgement, its operand is large, and detection efficiency is low.
The second picture analogies degree detection technique: MD5(message-digest algorithm5, message digest algorithm) code calculation detects.
Two pictures are calculated respectively to the MD5 code of all images content, if identical, represent that this two pictures is identical picture; Otherwise, represent that this two pictures is not identical picture.
The efficiency that MD5 code calculation detects is higher, but can only detect living picture, picture for the little change in part (such as increasing local word, pattern and watermark etc.) but can detect out, for the reusable detection case of picture in above-mentioned website inapplicable.
Therefore, need to find at present a kind of suitable picture analogies degree detection method, can detect exactly reusable picture in website, or only reusable picture after little change be made in part.
Summary of the invention
The embodiment of the present application provides a kind of similarity detection method and equipment of picture, not high in order to solve the similarity detection accuracy of current picture, or inspection does not measure the problem of only part being made the similar pictures after little change.
A similarity detection method for picture, described method comprises:
Target Photo is carried out after cutting according at least one slit mode, obtain a plurality of Target Photo pieces;
For each the Target Photo piece obtaining, carry out respectively following operation:
Determine the slit mode of the Target Photo piece obtaining, and this Target Photo piece position in Target Photo;
From the picture block to be measured of having stored, determine at least one picture block to be measured, wherein, the picture block to be measured of determining is to obtain with the slit mode identical with this Target Photo piece, and the position of the picture block to be measured of determining in picture to be measured is identical with this Target Photo piece position in Target Photo;
The MD5 code of this Target Photo piece and the MD5 code of the picture block to be measured of determining are compared, and the picture to be measured under the identical picture block to be measured of MD5 code is as the picture similar to Target Photo.
A similarity checkout equipment for picture, described equipment comprises:
Cutting module, for Target Photo is carried out after cutting according at least one slit mode, obtains a plurality of Target Photo pieces;
Detection module, for carrying out respectively following operation for each the Target Photo piece obtaining:
Determine the slit mode of the Target Photo piece obtaining, and this Target Photo piece position in Target Photo, from the picture block to be measured of having stored, determine at least one picture block to be measured, wherein, the picture block to be measured of determining is to obtain with the slit mode identical with this Target Photo piece, and the position of the picture block to be measured of determining in picture to be measured is identical with this Target Photo piece position in Target Photo, and the MD5 code of this Target Photo piece and the MD5 code of the picture block to be measured of determining are compared, picture to be measured under the identical picture block to be measured of MD5 code is as the picture similar to Target Photo.
The application's beneficial effect is as follows:
The embodiment of the present application by picture to be measured and Target Photo all respectively cutting be a plurality of Target Photo pieces, similarity between picture is detected to the similarity being converted between Target Photo piece to be detected, if it is identical to carry out the MD5 code of two picture block of similarity detection, represent that Target Photo and the content of picture to be measured in same area are in full accord, can determine that Target Photo and picture to be measured are similar pictures, owing to having the content of subregion in full accord in two similar pictures determining, therefore, the accuracy of detection is higher; Meanwhile, due to without requiring the MD5 code of two pictures identical, but require the MD5 code of same area content wherein identical, can detect the local repetitive picture changing.
Accompanying drawing explanation
Fig. 1 is the method step schematic diagram of storing picture block to be measured in the present embodiment one;
Fig. 2 (a) and Fig. 2 (b) are optional picture slit mode in the present embodiment one;
Fig. 3 is the similarity testing process schematic diagram of picture in the embodiment of the present application two;
Fig. 4 is the similarity assay device structures schematic diagram of the picture in the embodiment of the present application three.
Embodiment
The embodiment of the present application by picture to be measured and Target Photo all respectively cutting be a plurality of Target Photo pieces, similarity between picture is detected to the similarity being converted between Target Photo piece to be detected, owing to carrying out two picture block of similarity detection, according to identical slit mode, carry out cutting and obtain, and position identical (i.e. two picture block that picture block is same area in picture) in the picture under separately, therefore, if it is identical to carry out the MD5 code of two picture block of similarity detection, represent that Target Photo and the content of picture to be measured in same area are in full accord, can determine that Target Photo and picture to be measured are similar pictures.
By the scheme of the embodiment of the present application, for current characteristic similarity detection technique, owing to determining in two similar pictures, there is the content of subregion in full accord, the accuracy of detection is higher; Meanwhile, for current MD5 code calculation detection technique, due to without requiring the MD5 code of two pictures identical, but require the MD5 code of same area content wherein identical, can detect the local repetitive picture changing.
Below in conjunction with Figure of description, the scheme of the embodiment of the present application is elaborated.
In the scheme of the embodiment of the present application, all pictures in website can be considered as to picture to be measured, take Target Photo as sample, from a large amount of pictures to be measured, find out reusable picture.To this, need first treat mapping sheet and carry out cutting, and calculate the MD5 code of each picture block to be measured after cutting, then the MD5 code of each picture block to be measured is stored in database, set up picture library to be measured.During detection, Target Photo is carried out to picture cutting according to the identical segmentation technique adopting when processing picture to be measured, obtain Target Photo cutting Target Photo piece afterwards, then calculate the MD5 code of each picture block, afterwards, the MD5 code of the picture block to be measured of storing in the MD5 code of each Target Photo piece and database is compared, carry out similarity detection.
Below by embodiment mono-, be described in the process of the MD5 code of storing picture block to be measured in database, and in embodiment bis-, describe and utilize the similarity of picture block to detect the process to the similarity testing goal of picture that reaches.
Embodiment mono-:
As shown in Figure 1, for storing the method step schematic diagram of picture block to be measured in application embodiment mono-, mainly comprise the following steps:
Step 101: determine multiple pictures to be measured.
If the present embodiment one is applied in detecting in shopping website under the scene of picture use repeatly of merchandising, the picture to be measured of determining in this step 101 can be in shopping website for showing each picture of merchandising, can from the whole network picture library of shopping website, determine picture to be measured.
Step 102: every picture to be measured is carried out respectively to cutting according at least one slit mode, obtain a plurality of picture block to be measured.
In the scheme of this step 102, picture to be measured can carry out cutting according to a kind of slit mode, also can carry out cutting according to multiple slit mode, when carrying out cutting according to multiple slit mode, the to be measured picture block of the sum total of the picture block to be measured obtaining after every kind of slit mode cutting for obtaining after this picture cutting to be measured.
In this step 102, can treat the cutting that mapping sheet carries out any-mode according to demand, for example, as shown in Figure 2 (a) shows, picture to be measured be carried out respectively obtaining after cutting 6 picture block to be measured according to following three kinds of slit modes:
Slit mode one: be 3 picture block to be measured by a picture transversally cutting to be measured, described 3 picture block to be measured from top to bottom ratio of area are 2:1:2.The advantage of slit mode one is: when picture to be measured is that certain position on the above-below direction at Target Photo has been done and modified operation (such as increasing local word, pattern and watermark etc.), as long as the content of some picture block to be measured does not change in these 3 picture block to be measured, the similar pictures that this picture to be measured is Target Photo just can be detected.
Slit mode two: 2 picture block to be measured that are equal areas by the longitudinal cutting of picture to be measured.The advantage of slit mode two is: when picture to be measured is that certain position on the left and right directions at Target Photo has been done and modified operation (such as increasing local word, pattern and watermark etc.), as long as the content of some picture block to be measured does not change in these 2 picture block to be measured, the similar pictures that this picture to be measured is Target Photo just can be detected.
Slit mode three: be syncopated as and overlap with center picture to be measured and area is 1 picture block to be measured of picture 1/4 to be measured.The advantage of slit mode three is: the picture block to be measured being syncopated as is positioned at the center of picture to be measured, due in most cases, the body matter of picture is positioned at the center position of picture, therefore, be syncopated as the body matter that has comprised picture in picture block to be measured, when body matter is identical, even if surrounding has carried out modifying operation to picture, or when picture surrounding is identical data (as white background), can both detect this picture to be measured and Target Photo is similar pictures, the situation that reduces erroneous judgement occurs.
Distinguishingly, can also be using the mode of non-divided picture to be measured as slit mode four, as shown in Fig. 2 (b), the mode detecting with the picture block to be measured after slit mode four cuttings is current MD5 code calculation detection technique.
It should be noted that, in this step 102, be not limited to above four kinds of slit modes, as by picture to be measured according to the slit mode of sphere of movements for the elephants cutting, be more than to represent intuitively optional slit mode by example.Every picture to be measured can select identical slit mode to carry out cutting, also can be different pictures to be measured and selects different slit modes.
Step 103: the picture block to be measured obtaining after all pictures to be measured are divided is divided set.
In the scheme of this step 103, for improving the follow-up detection efficiency that carries out picture block to be measured, in this step 103, can be divided in identity set divide the identical picture block to be measured in position that obtain and in picture to be measured according to same way.
For example, all pictures to be measured are all used three kinds of slit modes shown in Fig. 2 (a) to carry out cutting, after every picture cutting to be measured, obtain 6 picture block to be measured, can divide and obtain 6 set, set comprises the to be measured picture block 1 of picture to be measured after according to slit mode one cutting in 1, set comprises the to be measured picture block 2 of picture to be measured after according to slit mode one cutting in 2, by that analogy, comprises the to be measured picture block 6 of picture to be measured after according to slit mode three cuttings in set 6.
Step 104: the MD5 code of storing each picture block to be measured in each set.
Step 105: be each picture block allocation identification to be measured.
While detecting due to the similarity carrying out picture block, there is a picture block to be measured identical with the MD5 code of Target Photo piece, just determine the picture analogies under it separately, therefore, relevant between the sign of each picture block to be measured after same picture cutting to be measured and affiliated picture to be measured, the ID of each picture block to be measured that more preferably, can directly the ID of picture to be measured be obtained after this picture cutting to be measured.
While storing picture block to be measured in database, can be each picture block to be measured and distribute 2 fields, store respectively ID and the MD5 code of picture block to be measured, and set up the ID of picture block to be measured and the index relative between MD5 code, utilize this index relative can improve the speed of picture analogies degree testing process.
Scheme by above step 101 to step 105, realized the picture block to be measured that storage detects for similarity in database, because the picture to be measured in the whole network picture library changes greatly, therefore, in the scheme of the present embodiment one, also can carry out above-mentioned steps 101 to the operation of step 105 to the picture to be measured of real-time update.For example, while having newly-increased picture to be measured in the whole network picture library, ID and the MD5 code of the picture block to be measured after newly-increased picture cutting to be measured can be stored in database; When needing the deletion of mapping sheet, the picture block to be measured that can simultaneously delete this picture to be measured is stored in ID and the MD5 code in database.
After the ID of picture block to be measured and MD5 code are stored in database, can start to carry out the similarity testing process of picture.
Embodiment bis-:
As shown in Figure 3, the similarity testing process schematic diagram for picture in the present embodiment two, mainly comprises the following steps:
Step 201: Target Photo is carried out to cutting according at least one slit mode, obtain a plurality of Target Photo pieces.
In the scheme of this step 201, the Target Photo that can directly accept keeper's input carries out cutting operation, also can accept the Image ID of keeper's input, and utilize this Image ID to find out corresponding picture to be measured from the whole network picture library that comprises picture to be measured, and using this picture to be measured as Target Photo.
It should be noted that, can be identical to the slit mode of Target Photo and the slit mode for the treatment of mapping sheet in embodiment mono-, also can be incomplete same, as long as there is at least one slit mode, be the slit mode of Target Photo, be again the slit mode of picture to be measured.
For example, picture to be measured carries out cutting according to three kinds of modes shown in Fig. 2 (a), Target Photo can carry out cutting according to three kinds of modes shown in Fig. 2 (a), also can only adopt slit mode one to carry out cutting, also can carry out cutting according to four kinds of modes shown in Fig. 2 (b).
Step 202: select a Target Photo piece from described a plurality of Target Photo pieces.
Step 203: determine the slit mode of the Target Photo piece of selecting, and this Target Photo piece position in Target Photo.
Step 204: the picture block to be measured of selecting at least one and this Target Photo piece to carry out similarity detection from the picture block to be measured of having stored.
In order to guarantee that similarity between picture block detects, can reflect exactly the similarity between picture under it, therefore, two picture block of carrying out similarity detection should be to adopt identical slit mode and the identical picture block in position in picture.
For example, the Target Photo piece of selecting in step 202 is the Target Photo piece 1 of Target Photo after according to mode one cutting, and picture block to be measured is in the step 103 of embodiment mono-, according to three kinds of slit modes shown in Fig. 2 (a), be divided into 6 set, the to be measured picture block of the picture block to be measured of selecting in this step 202 for comprising in set 1.
Step 205: the MD5 code that extracts the picture block to be measured of selecting.
Suppose that database is opened up two its ID of field store and MD5 code for each picture block to be measured in embodiment mono-,, in this step 204, can directly from database, extract the MD5 code of the picture block to be measured of each selection.
Step 206: the MD5 code of Target Photo piece is compared with the MD5 code of the picture block to be measured of selecting respectively, determine the picture block to be measured that MD5 code is identical with the MD5 code of Target Photo piece.
The MD5 code of Target Photo piece can calculate any time after step 201 and before the execution of this step.
Step 207: according to the ID of the picture block to be measured of determining, determine the picture to be measured that it is affiliated, and using the picture to be measured of determining as the picture similar to Target Photo.
The picture block to be measured that the picture block to be measured selected in this step 202 of take comprises in set 1 is example, if the MD5 code of picture block to be measured that definite wherein ID is ABC and ABD is identical with the MD5 code of Target Photo piece, can determine that ID is that the picture to be measured of ABC and ABD is similar to Target Photo, exportable similar picture to be measured, or export the ID of similar picture to be measured, also can export similar picture to be measured and ID simultaneously.
Step 208: all whether executed is complete for Target Photo piece, if so, finishes; Otherwise, jump to step 202, until the similarity testing process of all Target Photo pieces is finished.
So far, utilize picture block to be measured after the picture cutting to be measured of storing in database and the similarity testing result of Target Photo piece to reflect that the scheme of picture analogies degree finishes.
Embodiment tri-:
As shown in Figure 4, similarity assay device structures schematic diagram for picture in the present embodiment three, described similarity checkout equipment can be stored picture block to be measured according to the mode of embodiment mono-in database, and then according to the mode of embodiment bis-, carries out the similarity detection of picture, is illustrated respectively below:
Similarity checkout equipment comprises cutting module 11 and memory module 12, wherein: cutting module 11 can have the ability of access the whole network picture library, from the whole network picture library, extract picture to be measured, and multiple pictures to be measured that extract are carried out respectively to cutting according at least one slit mode, after every picture cutting to be measured, obtain a plurality of picture block to be measured, slit mode as shown according to Fig. 2 (b) is treated mapping sheet and is carried out cutting, and every picture cutting to be measured is 7 picture block to be measured.
Memory module 12 is for the picture block all to be measured obtaining is divided to set, and the MD5 code of each picture block to be measured of obtaining of storage, and wherein, the picture block to be measured in identity set obtains according to identical slit mode, and the position in picture to be measured is identical.For example, treat mapping sheet carry out after cutting according to the slit mode shown in Fig. 2 (b) when cutting module, memory module 12 can be divided into 7 set, and in each set, the dividing mode of the picture block to be measured of storage is identical with the position in picture to be measured.
Described memory module 12 is also for the sign of each picture block to be measured that the sign of picture to be measured is obtained after this picture cutting to be measured, particularly, memory module 12 can be each picture block to be measured and opens up two fields, the sign of one of them field store picture block to be measured, another is the MD5 code of this picture block to be measured of field store one to one.
When the similarity checkout equipment of picture has been stored after picture block to be measured in this locality, cutting module 11 is also for by the input port outwards providing, the Target Photo that receiving management person sends or the ID of Target Photo, when receive be the ID of Target Photo time, can from the whole network picture library, extract Target Photo, and the Target Photo of extraction is carried out respectively after cutting according at least one slit mode, obtain a plurality of Target Photo pieces, slit mode as shown according to Fig. 2 (b) carries out cutting to Target Photo, obtains 7 Target Photo pieces.
The similarity checkout equipment of described picture also comprises:
Detection module 13, for carrying out after cutting at 11 pairs of Target Photos of cutting module, for each the Target Photo piece obtaining, carry out respectively following operation:
Determine the slit mode of the Target Photo piece obtaining, and this Target Photo piece position in Target Photo, from the picture block to be measured of having stored, determine at least one picture block to be measured, wherein, the picture block to be measured of determining is to obtain with the slit mode identical with this Target Photo piece, and the position of the picture block to be measured of determining in picture to be measured is identical with this Target Photo piece position in Target Photo, and the MD5 code of this Target Photo piece and the MD5 code of the picture block to be measured of determining are compared, picture to be measured under the identical picture block to be measured of MD5 code is as the picture similar to Target Photo.
Owing to having stored the sign of picture block to be measured in memory module 12, and the sign of the picture to be measured that this sign is affiliated with picture block to be measured is identical, therefore, detection module 13 is specifically for after comparing the MD5 code of the MD5 code of Target Photo piece and picture block to be measured, determine the sign of the picture block to be measured that MD5 code is identical with the MD5 code of Target Photo piece, and determine the picture to be measured under picture block to be measured according to the sign of picture block to be measured, and then using this picture to be measured as the picture similar to Target Photo.
The scheme providing by the embodiment of the present application, on the higher basis of the accuracy that can detect in similarity, can also detect the multiimage that there is less change part, and the reliability of detection is high.
Those skilled in the art should understand, the application's embodiment can be provided as method, system or computer program.Therefore, the application can adopt complete hardware implementation example, implement software example or in conjunction with the form of the embodiment of software and hardware aspect completely.And the application can adopt the form that wherein includes the upper computer program of implementing of computer-usable storage medium (including but not limited to magnetic disk memory, CD-ROM, optical memory etc.) of computer usable program code one or more.
The application is with reference to describing according to process flow diagram and/or the block scheme of the method for the embodiment of the present application, equipment (system) and computer program.Should understand can be in computer program instructions realization flow figure and/or block scheme each flow process and/or the flow process in square frame and process flow diagram and/or block scheme and/or the combination of square frame.Can provide these computer program instructions to the processor of multi-purpose computer, special purpose computer, Embedded Processor or other programmable data processing device to produce a machine, the instruction of carrying out by the processor of computing machine or other programmable data processing device is produced for realizing the device in the function of flow process of process flow diagram or a plurality of flow process and/or square frame of block scheme or a plurality of square frame appointments.
These computer program instructions also can be stored in energy vectoring computer or the computer-readable memory of other programmable data processing device with ad hoc fashion work, the instruction that makes to be stored in this computer-readable memory produces the manufacture that comprises command device, and this command device is realized the function of appointment in flow process of process flow diagram or a plurality of flow process and/or square frame of block scheme or a plurality of square frame.
These computer program instructions also can be loaded in computing machine or other programmable data processing device, make to carry out sequence of operations step to produce computer implemented processing on computing machine or other programmable devices, thereby the instruction of carrying out is provided for realizing the step of the function of appointment in flow process of process flow diagram or a plurality of flow process and/or square frame of block scheme or a plurality of square frame on computing machine or other programmable devices.
In a typical configuration, described computer equipment comprises one or more processors (CPU), input/output interface, network interface and internal memory.Internal memory may comprise the volatile memory in computer-readable medium, and the forms such as random access memory (RAM) and/or Nonvolatile memory, as ROM (read-only memory) (ROM) or flash memory (flash RAM).Internal memory is the example of computer-readable medium.Computer-readable medium comprises that permanent and impermanency, removable and non-removable media can realize information by any method or technology and store.Information can be module or other data of computer-readable instruction, data structure, program.The example of the storage medium of computing machine comprises, but be not limited to phase transition internal memory (PRAM), static RAM (SRAM), dynamic RAM (DRAM), the random access memory of other types (RAM), ROM (read-only memory) (ROM), Electrically Erasable Read Only Memory (EEPROM), fast flash memory bank or other memory techniques, read-only optical disc ROM (read-only memory) (CD-ROM), digital versatile disc (DVD) or other optical memory, magnetic magnetic tape cassette, the storage of tape magnetic rigid disk or other magnetic storage apparatus or any other non-transmission medium, can be used for the information that storage can be accessed by computing equipment.According to defining herein, computer-readable medium does not comprise the computer readable media (transitory media) of non-standing, as data-signal and the carrier wave of modulation.
Although described the application's preferred embodiment, once those skilled in the art obtain the basic creative concept of cicada, can make other change and modification to these embodiment.So claims are intended to all changes and the modification that are interpreted as comprising preferred embodiment and fall into the application's scope.
Obviously, those skilled in the art can carry out various changes and modification and the spirit and scope that do not depart from the application to the application.Like this, if within these of the application are revised and modification belongs to the scope of the application's claim and equivalent technologies thereof, the application is also intended to comprise these changes and modification interior.
Claims (10)
1. a similarity detection method for picture, is characterized in that, described method comprises:
Target Photo is carried out after cutting according at least one slit mode, obtain a plurality of Target Photo pieces;
For each the Target Photo piece obtaining, carry out respectively following operation:
Determine the slit mode of the Target Photo piece obtaining, and this Target Photo piece position in Target Photo;
From the picture block to be measured of having stored, determine at least one picture block to be measured, wherein, the picture block to be measured of determining is to obtain with the slit mode identical with this Target Photo piece, and the position of the picture block to be measured of determining in picture to be measured is identical with this Target Photo piece position in Target Photo;
The MD5 code of this Target Photo piece and the MD5 code of the picture block to be measured of determining are compared, and the picture to be measured under the identical picture block to be measured of MD5 code is as the picture similar to Target Photo.
2. the method for claim 1, is characterized in that, stores in the following manner picture block to be measured:
Multiple pictures to be measured are carried out respectively to cutting according at least one slit mode, after every picture cutting to be measured, obtain a plurality of picture block to be measured;
The picture block all to be measured obtaining is divided to set, and store the MD5 code of each picture block to be measured obtaining, wherein, the picture block to be measured in identity set obtains according to identical slit mode, and the position in picture to be measured is identical.
3. method as claimed in claim 2, is characterized in that, determines and specifically comprise at least one picture block to be measured from the picture block to be measured of having stored:
According to the slit mode of Target Photo piece and this Target Photo piece position in Target Photo, determine a set that comprises picture block to be measured of dividing, wherein, the slit mode of the picture block to be measured comprising in this set slit mode fast with Target Photo is identical, and the position in picture to be measured is identical with the position of Target Photo piece in Target Photo;
The picture block to be measured comprising in definite set at least one picture block to be measured for determining from the picture block to be measured of having stored.
4. method as claimed in claim 2, is characterized in that, described method also comprises:
The sign of each picture block to be measured that the sign of picture to be measured is obtained after this picture cutting to be measured;
Picture to be measured under the identical picture block to be measured of MD5 code, as the picture similar to Target Photo, specifically comprises:
Determine the sign of the picture block to be measured that MD5 code is identical with the MD5 code of Target Photo piece;
According to the sign of picture block to be measured, determine the picture to be measured under picture block to be measured, and using this picture to be measured as the picture similar to Target Photo.
5. method as claimed in claim 2, is characterized in that, having at least one slit mode is Target Photo to be carried out to the slit mode of cutting, is again to treat the slit mode that mapping sheet carries out cutting.
6. method as claimed in claim 2, is characterized in that, every picture to be measured is carried out respectively to cutting according to following three kinds of slit modes, obtains 6 picture block to be measured after every picture cutting to be measured:
Slit mode one: be 3 picture block to be measured by picture transversally cutting to be measured, described 3 picture block to be measured from top to bottom ratio of area are 2:1:2;
Slit mode two: 2 picture block to be measured that are equal areas by the longitudinal cutting of picture to be measured;
Slit mode three: be syncopated as and overlap with center picture to be measured and area is 1 picture block to be measured of picture 1/4 to be measured;
The picture block all to be measured obtaining is divided into 6 set, and wherein, the picture block to be measured in identity set obtains according to identical slit mode, and the position in picture to be measured is identical.
7. a similarity checkout equipment for picture, is characterized in that, described equipment comprises:
Cutting module, for Target Photo is carried out after cutting according at least one slit mode, obtains a plurality of Target Photo pieces;
Detection module, for carrying out respectively following operation for each the Target Photo piece obtaining:
Determine the slit mode of the Target Photo piece obtaining, and this Target Photo piece position in Target Photo, from the picture block to be measured of having stored, determine at least one picture block to be measured, wherein, the picture block to be measured of determining is to obtain with the slit mode identical with this Target Photo piece, and the position of the picture block to be measured of determining in picture to be measured is identical with this Target Photo piece position in Target Photo, and the MD5 code of this Target Photo piece and the MD5 code of the picture block to be measured of determining are compared, picture to be measured under the identical picture block to be measured of MD5 code is as the picture similar to Target Photo.
8. equipment as claimed in claim 7, is characterized in that,
Cutting module, also for multiple pictures to be measured are carried out respectively to cutting according at least one slit mode, obtains a plurality of picture block to be measured after every picture cutting to be measured;
Described equipment also comprises:
Memory module, for the picture block all to be measured obtaining is divided to set, and stores the MD5 code of each picture block to be measured obtaining, and wherein, the picture block to be measured in identity set obtains according to identical slit mode, and the position in picture to be measured is identical.
9. equipment as claimed in claim 8, is characterized in that,
Memory module, also for the sign of each picture block to be measured that the sign of picture to be measured is obtained after this picture cutting to be measured;
Detection module, specifically for carrying out respectively following operation for each the Target Photo piece obtaining:
Determine the slit mode of the Target Photo piece obtaining, and this Target Photo piece position in Target Photo, from the picture block to be measured of having stored, determine at least one picture block to be measured, wherein, the picture block to be measured of determining is to obtain with the slit mode identical with this Target Photo piece, and the position of the picture block to be measured of determining in picture to be measured is identical with this Target Photo piece position in Target Photo, and the MD5 code of this Target Photo piece and the MD5 code of the picture block to be measured of determining are compared, determine the sign of the picture block to be measured that MD5 code is identical with the MD5 code of Target Photo piece, and determine the picture to be measured under picture block to be measured according to the sign of picture block to be measured, using this picture to be measured as the picture similar to Target Photo.
10. equipment as claimed in claim 8, is characterized in that,
Cutting module, specifically for every picture to be measured is carried out respectively to cutting according to following three kinds of slit modes, obtains 6 picture block to be measured after every picture cutting to be measured:
Slit mode one: be 3 picture block to be measured by picture transversally cutting to be measured, described 3 picture block to be measured from top to bottom ratio of area are 2:1:2;
Slit mode two: 2 picture block to be measured that are equal areas by the longitudinal cutting of picture to be measured;
Slit mode three: be syncopated as and overlap with center picture to be measured and area is 1 picture block to be measured of picture 1/4 to be measured;
Memory module, specifically for the picture block all to be measured obtaining is divided into 6 set, wherein, the picture block to be measured in identity set obtains according to identical slit mode, and the position in picture to be measured is identical.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310140673.2A CN104112284B (en) | 2013-04-22 | 2013-04-22 | The similarity detection method and equipment of a kind of picture |
| HK15102139.1A HK1201627A1 (en) | 2013-04-22 | 2015-03-03 | Method for detecting image similarity and device thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310140673.2A CN104112284B (en) | 2013-04-22 | 2013-04-22 | The similarity detection method and equipment of a kind of picture |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104112284A true CN104112284A (en) | 2014-10-22 |
| CN104112284B CN104112284B (en) | 2017-10-13 |
Family
ID=51709061
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310140673.2A Active CN104112284B (en) | 2013-04-22 | 2013-04-22 | The similarity detection method and equipment of a kind of picture |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN104112284B (en) |
| HK (1) | HK1201627A1 (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105404696A (en) * | 2015-12-23 | 2016-03-16 | 北京奇虎科技有限公司 | Method, system and device for downloading photographs in photograph album |
| CN105512328A (en) * | 2015-12-23 | 2016-04-20 | 北京奇虎科技有限公司 | Method, system and device for realizing uploading of album photos |
| CN105872760A (en) * | 2015-12-02 | 2016-08-17 | 乐视网信息技术（北京）股份有限公司 | Video play monitoring method and device |
| CN107729935A (en) * | 2017-10-12 | 2018-02-23 | 杭州贝购科技有限公司 | The recognition methods of similar pictures and device, server, storage medium |
| CN109284894A (en) * | 2018-08-10 | 2019-01-29 | 广州虎牙信息科技有限公司 | Picture examination method, apparatus, storage medium and computer equipment |
| CN110099237A (en) * | 2018-01-31 | 2019-08-06 | 腾讯科技（深圳）有限公司 | Image processing method, electronic device and computer readable storage medium |
| CN110378750A (en) * | 2019-07-25 | 2019-10-25 | 秒针信息技术有限公司 | Image rendering method, device, equipment and storage medium |
| WO2020155488A1 (en) * | 2019-01-31 | 2020-08-06 | 平安科技（深圳）有限公司 | Picture duplicate checking method and apparatus, computer device and storage medium |
| CN112529111A (en) * | 2020-12-28 | 2021-03-19 | 广东国粒教育技术有限公司 | Method for calculating class preparation innovation degree of teacher based on ppt document comparison technology |
| CN113066121A (en) * | 2019-12-31 | 2021-07-02 | 深圳迈瑞生物医疗电子股份有限公司 | Image analysis system and method for identifying repeat cells |
| CN113507485A (en) * | 2021-08-12 | 2021-10-15 | 河北民族师范学院 | Cloud security access system and method |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2002037331A1 (en) * | 1999-10-19 | 2002-05-10 | Microsoft Corporation | System and method for hashing digital images |
| CN101576932A (en) * | 2009-06-16 | 2009-11-11 | 阿里巴巴集团控股有限公司 | Close-repetitive picture computer searching method and device |
| CN101895789A (en) * | 2010-08-09 | 2010-11-24 | 北京海尔集成电路设计有限公司 | Method and device for detecting duplicate contents in television signals |
| CN102122389A (en) * | 2010-01-12 | 2011-07-13 | 阿里巴巴集团控股有限公司 | Method and device for judging image similarity |
| CN102270336A (en) * | 2011-07-06 | 2011-12-07 | 北京航空航天大学 | Safe fragile watermarking method based on multiple dependency structures |
| US20120096564A1 (en) * | 2010-10-13 | 2012-04-19 | Sony Corporation | Data integrity protecting and verifying methods, apparatuses and systems |
| CN102521838A (en) * | 2011-12-19 | 2012-06-27 | 国家计算机网络与信息安全管理中心 | Image searching/matching method and system for the same |
| CN102737254A (en) * | 2012-06-15 | 2012-10-17 | 常州南京大学高新技术研究院 | Identification method of mark image |
| CN102880726A (en) * | 2012-10-23 | 2013-01-16 | 深圳市宜搜科技发展有限公司 | Image filter method and image filter system |
-
2013
- 2013-04-22 CN CN201310140673.2A patent/CN104112284B/en active Active
-
2015
- 2015-03-03 HK HK15102139.1A patent/HK1201627A1/en unknown
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2002037331A1 (en) * | 1999-10-19 | 2002-05-10 | Microsoft Corporation | System and method for hashing digital images |
| CN101576932A (en) * | 2009-06-16 | 2009-11-11 | 阿里巴巴集团控股有限公司 | Close-repetitive picture computer searching method and device |
| CN102122389A (en) * | 2010-01-12 | 2011-07-13 | 阿里巴巴集团控股有限公司 | Method and device for judging image similarity |
| CN101895789A (en) * | 2010-08-09 | 2010-11-24 | 北京海尔集成电路设计有限公司 | Method and device for detecting duplicate contents in television signals |
| US20120096564A1 (en) * | 2010-10-13 | 2012-04-19 | Sony Corporation | Data integrity protecting and verifying methods, apparatuses and systems |
| CN102270336A (en) * | 2011-07-06 | 2011-12-07 | 北京航空航天大学 | Safe fragile watermarking method based on multiple dependency structures |
| CN102521838A (en) * | 2011-12-19 | 2012-06-27 | 国家计算机网络与信息安全管理中心 | Image searching/matching method and system for the same |
| CN102737254A (en) * | 2012-06-15 | 2012-10-17 | 常州南京大学高新技术研究院 | Identification method of mark image |
| CN102880726A (en) * | 2012-10-23 | 2013-01-16 | 深圳市宜搜科技发展有限公司 | Image filter method and image filter system |
Non-Patent Citations (3)
| Title |
|---|
| DCHZVK: ""谷歌的图片搜索原理"", 《百度文库》 * |
| 夏彬: ""基于内容的近似图像检测算法研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 张斌: ""基于感知哈希与数字水印图像内容认证技术研究"", 《中国博士学位论文全文数据库信息科技辑》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105872760A (en) * | 2015-12-02 | 2016-08-17 | 乐视网信息技术（北京）股份有限公司 | Video play monitoring method and device |
| CN105512328A (en) * | 2015-12-23 | 2016-04-20 | 北京奇虎科技有限公司 | Method, system and device for realizing uploading of album photos |
| CN105404696A (en) * | 2015-12-23 | 2016-03-16 | 北京奇虎科技有限公司 | Method, system and device for downloading photographs in photograph album |
| CN107729935A (en) * | 2017-10-12 | 2018-02-23 | 杭州贝购科技有限公司 | The recognition methods of similar pictures and device, server, storage medium |
| CN107729935B (en) * | 2017-10-12 | 2019-11-12 | 杭州贝购科技有限公司 | The recognition methods of similar pictures and device, server, storage medium |
| CN110099237B (en) * | 2018-01-31 | 2021-08-17 | 腾讯科技（深圳）有限公司 | Image processing method, electronic device, and computer-readable storage medium |
| CN110099237A (en) * | 2018-01-31 | 2019-08-06 | 腾讯科技（深圳）有限公司 | Image processing method, electronic device and computer readable storage medium |
| CN109284894A (en) * | 2018-08-10 | 2019-01-29 | 广州虎牙信息科技有限公司 | Picture examination method, apparatus, storage medium and computer equipment |
| WO2020155488A1 (en) * | 2019-01-31 | 2020-08-06 | 平安科技（深圳）有限公司 | Picture duplicate checking method and apparatus, computer device and storage medium |
| CN110378750A (en) * | 2019-07-25 | 2019-10-25 | 秒针信息技术有限公司 | Image rendering method, device, equipment and storage medium |
| CN113066121A (en) * | 2019-12-31 | 2021-07-02 | 深圳迈瑞生物医疗电子股份有限公司 | Image analysis system and method for identifying repeat cells |
| CN112529111A (en) * | 2020-12-28 | 2021-03-19 | 广东国粒教育技术有限公司 | Method for calculating class preparation innovation degree of teacher based on ppt document comparison technology |
| CN113507485A (en) * | 2021-08-12 | 2021-10-15 | 河北民族师范学院 | Cloud security access system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104112284B (en) | 2017-10-13 |
| HK1201627A1 (en) | 2015-09-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104112284A (en) | Method and equipment for detecting similarity of images | |
| JP6607061B2 (en) | Information processing apparatus, data comparison method, and data comparison program | |
| CN111008620A (en) | Target user identification method and device, storage medium and electronic equipment | |
| Kumar et al. | SOMES: an efficient SOM technique for event summarization in multi-view surveillance videos | |
| CN111159697A (en) | Key detection method and device and electronic equipment | |
| KR20160085004A (en) | Duplication Image File Searching Method and Apparatus | |
| CN111224923A (en) | Detection method, device and system for counterfeit websites | |
| CN111858929A (en) | Network crawler detection method, system and device based on graph neural network | |
| CN112579623A (en) | Method, device, storage medium and equipment for storing data | |
| CN109961061A (en) | A kind of edge calculations video data structure method and system | |
| EP3821366A1 (en) | Systems, methods, and computer-readable media for improved table identification using a neural network | |
| CN108664900B (en) | Method and equipment for identifying similarities and differences of written works | |
| US9317125B2 (en) | Searching of line pattern representations using gestures | |
| CN104965853A (en) | Method and system for recommending aggregation application, method and device for aggregating various recommendation resources | |
| Tauheed et al. | Configuring spatial grids for efficient main memory joins | |
| Xiao et al. | Confidence map based 3D cost aggregation with multiple minimum spanning trees for stereo matching | |
| US20220318359A1 (en) | Method and apparatus for deep learning-based real-time on-device authentication | |
| CN112949736B (en) | Feature matching method and related equipment | |
| Li et al. | Hashing-based approximate DBSCAN | |
| CN111241893A (en) | Identification recognition method, device and system | |
| Li et al. | Community Detection Using Revised Medoid-Shift Based on KNN | |
| CN113065071B (en) | Product information recommendation method and computer equipment | |
| Agasiev | Generalized information content based on variability map for exploratory landscape analysis of global optimization problems | |
| US20240161365A1 (en) | Enhancing images in text documents | |
| CN111124144B (en) | Input data processing method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 1201627 Country of ref document: HK |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: GR Ref document number: 1201627 Country of ref document: HK |
|
| TR01 | Transfer of patent right |
Effective date of registration: 20240221 Address after: Singapore Patentee after: Alibaba Singapore Holdings Ltd. Country or region after: Singapore Address before: Cayman Islands Grand Cayman capital building, a four storey No. 847 mailbox Patentee before: ALIBABA GROUP HOLDING Ltd. Country or region before: Cayman Islands |
|
| TR01 | Transfer of patent right |