CN100462978C - Image searching method and system - Google Patents
Image searching method and system Download PDFInfo
- Publication number
- CN100462978C CN100462978C CNB2007100984941A CN200710098494A CN100462978C CN 100462978 C CN100462978 C CN 100462978C CN B2007100984941 A CNB2007100984941 A CN B2007100984941A CN 200710098494 A CN200710098494 A CN 200710098494A CN 100462978 C CN100462978 C CN 100462978C
- Authority
- CN
- China
- Prior art keywords
- image
- similarity
- master sample
- sample image
- collection
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Processing Or Creating Images (AREA)
Abstract
An image indexing system is prepared as using the first similarity set to compare image requiring to be indexed with each standard sample image in standard sample image set, indexing image to be indexed in image library according to the first similarity set and similarity set library and forming similarity set library by each image to be indexed in image library and the second similarity set of each standard sample image in standard sample image set.
Description
Technical field
The present invention relates to the data retrieval field, particularly a kind of image search method and system.
Background technology
The research of relevant image retrieval has just begun since the seventies in 20th century, a kind of image retrieval technologies (TBIR that is based on the image text index, Text-based Image Retrieval), also be current application the most extensively, the most ripe a kind of image search method, mainly be to utilize the image text mentioned way to describe the feature of image, as picture specification, source etc.Another kind is based on the image retrieval (CBIR of content, Content-based Image Retrieval) technology, it is representative image retrieval technique development in future direction, mainly be contents semantic to image, the image retrieval technologies of analyzing and retrieving as the color of image, texture, layout etc.Along with developing rapidly of the computer technology and the communication technology, the multi-medium data that comprises digital picture of retrieving and browse magnanimity becomes urgent day by day problem.
The inventor notices in the invention process: for the image retrieval technologies of overwhelming majority's employing based on the image text index, there is following deficiency in it:
1) text index itself does not have objective unified standard, has certain subjectivity, for example may be different to the text message of different " people " possibility index of same width of cloth image;
2) by image text index information retrieval, may cause the retrieval failure owing to the image text index in the image library of text key word of retrieving and retrieval is inconsistent, may there be a large amount of incoherent result for retrieval in the result for retrieval simultaneously.That is,, so do not exist inevitable relatedly between label and the picture material, can not retrieve the true content of image because the relation between label and the picture material is artificial setting.
For present CBIR technology, the general collection and the processing image resource, features such as extraction such as color characteristic, textural characteristics, shape facility are as the image data base image index, calculate the similarity size of the image in retrieving images and the image library then, extract satisfy threshold value record as a result of, export according to the mode of similarity descending.But the inventor notices that it has the following disadvantages:
1) because the employing color histogram comes the method for the similarity between computed image fairly simple, so it can not reflect the space characteristics of object in the image;
2) because texture description is relatively more difficult, the sample query mode is all adopted in the retrieval of texture, repeatedly mutual, progressively refinement, thus cause inquiry velocity slow;
3) when employing was retrieved based on shape facility, shape or the profile of user by sketching the contours image detected the similar image of shape from image library.Search method based on this feature has two kinds: i) split image obtains the outline line of target image through behind the edge extracting, retrieves at the shape facility that this outline line carries out.Ii) directly seek suitable vector characteristic and be used for searching algorithm at figure.Obviously, it is more complicated to handle this structuring retrieval, need do more pre-service or the like, makes that this method is not only slow but also realize very difficult.
As seen by above-mentioned, above technology exists: it is accurate inadequately 1) to retrieve the Query Result that returns; Perhaps 2) calculation of complex needs repeatedly alternately, and inquiry velocity is slow.Thereby cause the conventional images retrieval technique not in actual widespread use.
Summary of the invention
The invention provides a kind of image search method and system, device, in order to solve exist in the prior art can not satisfy retrieval when the image retrieval simultaneously efficiently and result for retrieval problem accurately.
The invention provides a kind of image search method, comprise the steps:
Relatively need the first similarity collection of retrieving images and each master sample image of master sample image set;
Similarity height according to second similarity collection in the similarity collection storehouse and the described first similarity collection, retrieve the image to be retrieved of answering with the described second similarity set pair from image library, described similarity Ji Ku is the similarity Ji Ku that the second similarity collection of each image to be retrieved in the described image library and described each master sample image of master sample image set is formed.
Preferably, further comprise the steps:
When image library is added image to be retrieved, the image to be retrieved of described interpolation and the second similarity collection of described each master sample image of master sample image set are added into similarity Ji Ku.
Preferably, further comprise the steps:
With the image image that similarity is little each other as the master sample image in the described master sample image set.
Preferably, described similarity is to come comparison according to the proper vector between the image.
The present invention also provides a kind of image indexing system, comprising:
First comparison module relatively needing to be used for the first similarity collection of retrieving images and each master sample image of master sample image set;
Second comparison module is used for each image to be retrieved in movement images storehouse and the second similarity collection of described each master sample image of master sample image set;
Memory module links to each other with described second comparison module, is used to store the similarity Ji Ku that the described second comparison module second similarity collection is relatively formed;
Retrieval module, link to each other with first comparison module, memory module, be used for according to the similarity of the second similarity collection in the described similarity collection storehouse of the described first comparison module described first similarity collection relatively and described memory module just from image library, retrieving the image to be retrieved of answering with the described second similarity set pair.
Preferably, further comprise:
Output module links to each other with retrieval module, is used for exporting the image to be retrieved that described retrieval module retrieval retrieves from image library.
Preferably, further comprise:
Add module, link to each other with second comparison module, memory module, be used for when when image library is added image to be retrieved, the second similarity collection of the image to be retrieved of the described interpolation of the described second comparison module comparison and described each master sample image of master sample image set be added into the similarity Ji Ku of described memory module.
Preferably, further comprise:
The master sample image is chosen module, is used for the master sample image of the image image that similarity is little each other as described master sample image set.
Preferably, further comprise:
The proper vector load module is used for the proper vector of input picture;
Described system comes retrieving images according to the described similarity between the proper vector of image.
Beneficial effect of the present invention is as follows:
Because each image to be retrieved in the present invention elder generation movement images storehouse and the second similarity collection of described each master sample image of master sample image set; And the similarity Ji Ku of the second similarity collection that will compare composition; The pre-determined master sample image set of this employing is as the characteristics of image scale of measurement, thereby has stronger objectivity, consistance;
Because after having set up similarity Ji Ku, by relatively needing the first similarity collection of retrieving images and each master sample image of master sample image set; From image library, retrieve image to be retrieved according to the first similarity collection and similarity Ji Ku again.Thereby by introducing master sample image set intermediary as a comparison, enlarged the feature space of image, comprised image feature information more, improved the precision of retrieval return results collection.Simultaneously, do not need the image in retrieving images and the image library is relatively calculated its similarity one by one, introduce master sample image set intermediary as a comparison but adopt, the similarity of each master sample image that is image in the good image library of calculated in advance and master sample image set is as its index vector, thereby improved retrieval rate.
Further, owing to adopted the method for definite characteristics of image vector to carry out the retrieval of similarity, it is simple, efficient, easy to implement to make invention implement.
Description of drawings
Fig. 1 is the synoptic diagram of image search method implementing procedure described in the embodiment of the invention;
Fig. 2 is an image indexing system structural representation described in the embodiment of the invention;
Fig. 3 is the synoptic diagram of image retrieval flow implementation described in the embodiment of the invention.
Embodiment
Below in conjunction with accompanying drawing the specific embodiment of the present invention is described.
The objective of the invention is in retrieving, to realize complicated at common image search method reality, retrieval rate is slower, the result set precision of returning is not good enough, thereby has proposed the search method that can improve retrieval performance, and takes this to propose corresponding implementation method.Realize simply, both improved the speed of retrieval, improved the precision effect of return results collection again.
Fig. 1 is an image search method implementing procedure synoptic diagram, as shown in the figure, comprises the steps: in the retrieval
In preferred the enforcement, when the choice criteria sample image, can be with the similarity between the image as basis, with the image image that similarity is little each other as the master sample image in the master sample image set, thereby make the proper vector of image in the similarity collection storehouse depict the feature of image more accurately, in other words, the proper vector of image can give expression to the mutual difference between the image in the similarity collection storehouse better in the similarity collection storehouse, thereby makes follow-up retrieval precision uprise.
By the preparation of above two steps, set up the similarity Ji Ku of image to be retrieved, just can be by in the past the retrieval of picture material having been become the retrieval of similarity in the later retrieving, thus avoided based on the direct drawback of retrieval of picture material.
Can also be in the enforcement if necessary in when image library is added image to be retrieved, also the image to be retrieved that added and the second similarity collection of each master sample image of master sample image set are added into similarity Ji Ku.
In preferred the enforcement, similarity height according to the second similarity collection and the described first similarity collection in the described similarity collection storehouse, from image library, retrieve the image to be retrieved of answering with the described second similarity set pair, because the image to be retrieved that comes out by similarity retrieval is not accurately to mate one by one, therefore when the output result for retrieval, can export by the height of similarity coupling according to the actual needs.
In the enforcement, similarity can just draw the height of similarity according to the coupling of the proper vector between the image, can optionally choose the quantity of characteristics of image vector, needs the precision height then to choose more proper vector.
Set forth based on the search method of image similarity comparison with an example more below and implement, may further comprise the steps:
1) determine that image library utilizes that characteristics of image vector obtains with the master sample image set in the similarity Ji Ku that forms according to similarity of each master sample image.
Among the embodiment, the S presentation video is the master sample image, [k] expression master sample picture numbers; The D presentation video is an image to be retrieved in the image library, [n] represents picture numbers to be retrieved, IS[i] similarity between [j] expression i master sample image and the j image to be retrieved, DX[i]=(IS[i] [1], IS[i] [2], ..., IS[i] [K]) be the similarity collection (the second similarity collection) of i master sample image.
Input master sample image (S[1], S[2] ..., S[K]) and the image library image (D[1], D[2] ..., D[n]), the image library image carries out the similarity comparison with the master sample image respectively one by one, calculates the similarity IS[i between them respectively] [j], as shown in the table:
So, obtain image library image D[i to be retrieved] relative standard's sample image (S[1], S[2] ..., S[K]) similar features vector set DX[i]=(IS[i] [1], IS[i] [2] ..., IS[i] [K]).
2) repetitive process 1) calculate and preserve all images storehouse image relative standard sample image (S[1], S[2] ..., S[K]) proper vector after, form similarity Ji Ku.
3) utilize and the retrieving images proper vector to determine similarity collection (the first similarity collection).
Input master sample image (S[1], S[2] ..., S[K]) and need retrieving images KEY, need retrieving images to carry out the similarity comparison with the master sample image respectively, calculate the preliminary similarity KS[j between them respectively], so, obtain needing retrieving images KEY relative standard sample image (S[1], S[2] ..., S[K]) the similar features vector similarity collection KEYX=(KS[1], KS[2] ..., KS[K]).
4) carrying out image retrieval calculates.
Calculate to need the vectorial DX[i of characteristics of image in retrieving images proper vector KEYX and the image library similarity collection storehouse respectively] between similarity FS[i] as its final similarity.Suppose preliminary similarity KS[j] and IS[i] span of [j] is 0-100, similarity FS[i] computing method can define like this, FS[i]=100-(ABS (KS[1]-IS[i] [1])+ABS (KS[2]-IS[i] [2])+...+ABS (KS[K]-IS[i] [K]))/K, the wherein absolute calculation of ABS (a) expression a.
In concrete the enforcement, the principle embodiment that needs similarity collection that retrieving images obtains and similarity Ji Ku to retrieve is a lot of based on utilizing, and therefore, present embodiment provides similarity FS[i] computing method but easily know this mode that is not limited in.
5) output Query Result.
FS[i] ordering in reverse order, top result is similar to search output result.
In the enforcement, before carrying out image retrieval, should determine good all images storehouse image relative standard sample image similar features vector earlier, the similar features vector that can utilize image library image relative standard sample image obtains the most similar retrieval with similarity between the similar features vector of retrieving images relative standard sample image and exports the result.
In order to strengthen retrieval effectiveness, can take the following ancillary method in preferred the enforcement:
1) choice criteria sample image collection.During the selection standard sample image, as the master sample image in the master sample image set, select to exist between those each images bigger difference as much as possible the image image that similarity is little each other as the master sample image.For example: can from different classes of images such as character image, landscape painting picture, respectively select one as sample image.
2) select preliminary similarity calculating method between image and the image.When selecting preliminary similarity calculating method, can select those computing method can reflect difference between the image, and have stronger objectivity and consistance.
The present invention also provides a kind of image indexing system, and Fig. 2 is the image indexing system structural representation, as shown in the figure, comprises in the image indexing system: first comparison module, second comparison module, memory module, retrieval module, wherein:
First comparison module relatively needing to be used for the first similarity collection of retrieving images and each master sample image of master sample image set;
Second comparison module is used for each image to be retrieved in movement images storehouse and the second similarity collection of described each master sample image of master sample image set;
Memory module links to each other with described second comparison module, is used to store the similarity Ji Ku that the described second comparison module second similarity collection is relatively formed;
Retrieval module links to each other with first comparison module, memory module, is used for retrieving image to be retrieved according to the described first similarity collection of described first comparison module comparison and the described similarity Ji Ku of described memory module from image library.
Can further include in preferred the enforcement:
Output module links to each other with retrieval module, is used for exporting described retrieval module and retrieves the image to be retrieved of answering with the described second similarity set pair from image library.
Add module, link to each other with second comparison module, memory module, be used for when when image library is added image to be retrieved, the second similarity collection of the image to be retrieved of the described interpolation of the described second comparison module comparison and described each master sample image of master sample image set be added into the similarity Ji Ku of described memory module.
The master sample image is chosen module, is used for the master sample image of the image image that similarity is little each other as described master sample image set.
In preferred the enforcement, can further include:
The proper vector load module is used for the proper vector of input picture; Image indexing system comes retrieving images according to the described similarity between the proper vector of image.
Fig. 3 is an image retrieval flow implementation synoptic diagram, and briefly, as shown in the figure, image retrieval can be implemented by following basic procedure:
Step 301, newly-increased image library image to be retrieved;
Step 302, calculating proper vector to be retrieved are also added;
Step 303, image library characteristics of image vector to be retrieved;
Step 304, input need retrieving images;
Step 305, calculating need the retrieving images proper vector;
Step 306, utilize image library characteristics of image vector to carry out image retrieval to calculate;
Step 307, output Query Result.
Below in conjunction with example above-mentioned enforcement is described.
Suppose the image library that has one to have 1,000,000 images now.
At first, choose 5 master sample images (S[1], S[2] ..., S[5]) as the master sample image set.
Secondly, the similar features vector DX[i of 1,000,000 relative 5 the master sample images of image to be retrieved in difference computed image storehouse]=(IS[i] [1], IS[i] [2] ..., IS[i] [5]), i=1,2 ..., 1000000.
Once more, the similar features vector KEYX=of relative 5 the master sample images of calculating need retrieving images KEY (KS[1], KS[2] ..., KS[5]).
Then, calculate to need the vectorial DX[i of retrieving images proper vector KEYX and image library characteristics of image respectively] between final similarity FS[i], i=1,2 ..., 1000000.
At last, FS[i] (i=1,2 ..., 1000000) ordering in reverse order, top result (for example preceding 10) is similar to search output result.
By scheme of the present invention, can on the prior art basis, obtain fast speed and more accurate image retrieval effect very effectively.Simultaneously provided by the invention based on image similarity than retrieval scheme over the ground, difference that can be as requested can adopt different preliminary similarity calculating methods to obtain required retrieval effectiveness.Realize simply easy operating.
As seen from the above-described embodiment, the present invention: 1) proposed a kind of new image characteristic extracting method, adopted pre-determined master sample image set, had stronger objectivity, consistance as characteristics of image (vector) scale of measurement; 2) determine that the method for characteristics of image (vector) is fairly simple, do not need the image in retrieving images and the image library is relatively calculated its similarity one by one, introduce master sample image set intermediary as a comparison but adopt, the similarity of each master sample image that is image in the good image library of calculated in advance and master sample image set has improved retrieval rate as its index vector; 3) owing to introduce master sample image set intermediary as a comparison, enlarge the feature space of image, comprised image feature information more, improved the precision of retrieval return results collection.
By above-mentioned enforcement also as can be known, because existing various image search method ubiquities are not high for the retrieval return results collection degree of accuracy in images with large data volume storehouse, speed is slow, be difficult to arrive application request.Therefore the present invention has proposed a kind of retrieval scheme based on the image similarity comparison on existing image search method basis.Adopt scheme of the present invention, can effectively improve the degree of accuracy of image retrieval return results collection, improve the retrieval rate of image simultaneously.The image similarity that can be applied in addition between two or more image libraries is mated automatically to be used as more retrieval utilization.
Obviously, those skilled in the art can carry out various changes and modification to the present invention and not break away from the spirit and scope of the present invention.Like this, if of the present invention these are revised and modification belongs within the scope of claim of the present invention and equivalent technologies thereof, then the present invention also is intended to comprise these changes and modification interior.
Claims (9)
1. an image search method is characterized in that, comprises the steps:
Relatively need the first similarity collection of retrieving images and each master sample image of master sample image set;
Similarity height according to second similarity collection in the similarity collection storehouse and the described first similarity collection, retrieve the image to be retrieved of answering with the described second similarity set pair from image library, described similarity Ji Ku is the similarity Ji Ku that the second similarity collection of each image to be retrieved in the described image library and described each master sample image of master sample image set is formed.
2. the method for claim 1 is characterized in that, further comprises the steps:
When image library is added image to be retrieved, the image to be retrieved of described interpolation and the second similarity collection of described each master sample image of master sample image set are added into similarity Ji Ku.
3. the method for claim 1 is characterized in that, further comprises the steps:
With the image image that similarity is little each other as the master sample image in the described master sample image set.
4. as the arbitrary described method of claim 1 to 3, it is characterized in that described similarity is to come comparison according to the proper vector between the image.
5. an image indexing system is characterized in that, comprising:
First comparison module relatively needing to be used for the first similarity collection of retrieving images and each master sample image of master sample image set;
Second comparison module is used for each image to be retrieved in movement images storehouse and the second similarity collection of described each master sample image of master sample image set;
Memory module links to each other with described second comparison module, is used to store the similarity Ji Ku that the described second comparison module second similarity collection is relatively formed;
Retrieval module, link to each other with first comparison module, memory module, be used for according to the similarity of the second similarity collection in the described similarity collection storehouse of the described first comparison module described first similarity collection relatively and described memory module just from image library, retrieving the image to be retrieved of answering with the described second similarity set pair.
6. system as claimed in claim 5 is characterized in that, further comprises:
Output module links to each other with retrieval module, is used for exporting the image to be retrieved that described retrieval module retrieves from image library.
7. system as claimed in claim 5 is characterized in that, further comprises:
Add module, link to each other with second comparison module, memory module, be used for when when image library is added image to be retrieved, the second similarity collection of the image to be retrieved of the described interpolation of the described second comparison module comparison and described each master sample image of master sample image set be added into the similarity Ji Ku of described memory module.
8. system as claimed in claim 5 is characterized in that, further comprises:
The master sample image is chosen module, is used for the master sample image of the image image that similarity is little each other as described master sample image set.
9. as the arbitrary described system of claim 5 to 8, it is characterized in that, further comprise:
The proper vector load module is used for the proper vector of input picture;
Described system comes retrieving images according to the described similarity between the proper vector of image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2007100984941A CN100462978C (en) | 2007-04-18 | 2007-04-18 | Image searching method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2007100984941A CN100462978C (en) | 2007-04-18 | 2007-04-18 | Image searching method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101030230A CN101030230A (en) | 2007-09-05 |
| CN100462978C true CN100462978C (en) | 2009-02-18 |
Family
ID=38715572
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2007100984941A Expired - Fee Related CN100462978C (en) | 2007-04-18 | 2007-04-18 | Image searching method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100462978C (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101739428B (en) * | 2008-11-10 | 2012-04-18 | 中国科学院计算技术研究所 | Method for establishing index for multimedia |
| CN101859326B (en) * | 2010-06-09 | 2012-04-18 | 南京大学 | Image searching method |
| CN102902826B (en) * | 2012-11-08 | 2016-07-06 | 公安部第三研究所 | A kind of image method for quickly retrieving based on reference picture index |
| CN106033443B (en) * | 2015-03-16 | 2019-05-07 | 北京大学 | A kind of expanding query method and device in vehicle retrieval |
| CN106528568A (en) * | 2015-09-11 | 2017-03-22 | 中兴通讯股份有限公司 | Method and system for information supply |
| US10405003B2 (en) * | 2017-01-20 | 2019-09-03 | Google Llc | Image compression based on semantic relevance |
| CN108805148B (en) * | 2017-04-28 | 2022-01-11 | 富士通株式会社 | Method of processing image and apparatus for processing image |
| CN112668632B (en) * | 2020-12-25 | 2022-04-08 | 浙江大华技术股份有限公司 | Data processing method and device, computer equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5586197A (en) * | 1993-09-02 | 1996-12-17 | Canon Kabushiki Kaisha | Image searching method and apparatus thereof using color information of an input image |
| US20020110276A1 (en) * | 2000-12-12 | 2002-08-15 | Nec Corporation | Image search system and image search method thereof |
| CN1551017A (en) * | 2003-05-08 | 2004-12-01 | ���ְ�˹��ʽ���� | Image searching device, method, programmme and storage medium storing said programme |
| US20050041863A1 (en) * | 2003-08-19 | 2005-02-24 | Ray Ajoy K. | Enabling content-based search of objects in an image database with reduced matching |
| CN1851703A (en) * | 2006-05-10 | 2006-10-25 | 南京大学 | Active semi-monitoring-related feedback method for digital image search |
-
2007
- 2007-04-18 CN CNB2007100984941A patent/CN100462978C/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5586197A (en) * | 1993-09-02 | 1996-12-17 | Canon Kabushiki Kaisha | Image searching method and apparatus thereof using color information of an input image |
| US20020110276A1 (en) * | 2000-12-12 | 2002-08-15 | Nec Corporation | Image search system and image search method thereof |
| CN1551017A (en) * | 2003-05-08 | 2004-12-01 | ���ְ�˹��ʽ���� | Image searching device, method, programmme and storage medium storing said programme |
| US20050041863A1 (en) * | 2003-08-19 | 2005-02-24 | Ray Ajoy K. | Enabling content-based search of objects in an image database with reduced matching |
| CN1851703A (en) * | 2006-05-10 | 2006-10-25 | 南京大学 | Active semi-monitoring-related feedback method for digital image search |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101030230A (en) | 2007-09-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100462978C (en) | Image searching method and system | |
| CN101201822B (en) | Method for searching visual lens based on contents | |
| CN112541490A (en) | Archive image information structured construction method and device based on deep learning | |
| US11157550B2 (en) | Image search based on feature values | |
| CN105677735A (en) | Video search method and apparatus | |
| CN102902826B (en) | A kind of image method for quickly retrieving based on reference picture index | |
| WO2006075902A1 (en) | Method and apparatus for category-based clustering using photographic region templates of digital photo | |
| Lee et al. | Video indexing: an approach based on moving object and track | |
| CN102236714A (en) | Extensible markup language (XML)-based interactive application multimedia information retrieval method | |
| CN113568995A (en) | Dynamic tile map making method based on retrieval conditions and tile map system | |
| CN113535788A (en) | Retrieval method, system, equipment and medium for marine environment data | |
| CN108491543A (en) | Image search method, image storage method and image indexing system | |
| CN113190645A (en) | Index structure establishing method, device, equipment and storage medium | |
| CN115455249A (en) | Double-engine driven multi-modal data retrieval method, equipment and system | |
| Truong et al. | Video search based on semantic extraction and locally regional object proposal | |
| Ahmad et al. | Indexing and retrieval of images by spatial constraints | |
| CN107423294A (en) | A kind of community image search method and system | |
| Matsui | Challenge for manga processing: Sketch-based manga retrieval | |
| CN103092935A (en) | Approximate copy image detection method based on scale invariant feature transform (SIFT) quantization | |
| Juan et al. | Content-based video retrieval system research | |
| CN104866818A (en) | Method and device for searching pictures including human faces | |
| Belarbi et al. | Indexing video by the content | |
| CN106776654B (en) | Data searching method and device | |
| US11500937B1 (en) | Data retrieval system | |
| Khawandi et al. | A survey on image indexing and retrieval based on content based image |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090218 Termination date: 20170418 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |