CN103955543A - Multimode-based clothing image retrieval method - Google Patents
Multimode-based clothing image retrieval method Download PDFInfo
- Publication number
- CN103955543A CN103955543A CN201410214929.4A CN201410214929A CN103955543A CN 103955543 A CN103955543 A CN 103955543A CN 201410214929 A CN201410214929 A CN 201410214929A CN 103955543 A CN103955543 A CN 103955543A
- Authority
- CN
- China
- Prior art keywords
- image
- clothing
- text
- candidate set
- retrieval
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
- G06F16/43—Querying
- G06F16/432—Query formulation
- G06F16/434—Query formulation using image data, e.g. images, photos, pictures taken by a user
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
- G06F16/48—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Library & Information Science (AREA)
- Mathematical Physics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Processing Or Creating Images (AREA)
Abstract
The invention provides a multimode-based clothing image retrieval method for comprehensive clothing text information and image information. The method comprises the steps of obtaining a candidate set through text retrieval and image multi-featured retrieval according to to-be-retrieved clothing image and text depiction which are input by a user, and then resetting the candidate set according to the feedback of the user so as to obtain a final retrieval result. According to the method provided by the invention, two-layer retrieval and a resetting process are built through descriptions of two modes of text and a clothing image to clothes, the recall ratio and the precision ratio of the clothing image retrieval are improved, and the retrieval efficiency is also improved.
Description
Technical field
The invention belongs to digital image processing techniques field, be specifically related to a kind of image of clothing search method based on multi-modal.
Background technology
Along with extensively popularizing of internet and developing rapidly of ecommerce, there is every day ten hundreds of merchandise news to pour in people's eyes, especially toggery information.How can fast and effectively from the data of magnanimity, search out the interested clothing information of people, to become problem in the urgent need to address.Clothing information comprises the text description of clothes and image appearance.Nowadays on the market, maximum search engines about costume retrieval is all based on text keyword, Taobao for example, Jingdone district, Amazon, ebay etc.Although these search engines are known and use by increasing user, the limitation of itself is obvious to all.When user wants to find certain a concrete clothes, can only describe by detailed key word, then from candidate's row of thousands of meters, filter out one by one own interested clothes, a large amount of operations that user carries out extremely test user's patience, have wasted user's plenty of time and energy.
In addition, image of clothing itself is also a kind of retrieval of content, and the image of clothing of usining obtains its same or similar clothing information as input.Similarly to scheme to search the system of figure, have: figure knows in Baidu, the picture searching of Google, TinEye, GazoPa etc., wherein CBIR is to scheme one of gordian technique of searching in figure.CBIR is by the extraction of characteristics of image and description, obtains identical or similar image retrieval.Yet traditional CBIR technology can not be carried out effective modeling to reach the target of accurate retrieval to this problem.
The basic performing step of existing CBIR technology is: (1) builds image library; (2) extract characteristics of image; (3), by distance (as Euclidean distance, manhatton distance etc.) between calculated characteristics, obtain the similarity degree of two width images; (4) user's input picture, obtains the image identical or similar with it.The method is applied in image of clothing searching system and has following problem: 1) ignored the text message that clothing information comprises; 2) between computed image, similarity need to be carried out the calculating of high dimension vector spacing, and the efficiency of linear sweep is extremely low; 3) lack mutual between user and system, lack the proof procedure to result for retrieval.
A kind of method of the image of clothing retrieval based on many features has been proposed in the image of clothing retrieval > > of the paper < < that the people such as the Hou Alin of Changchun Polytechnic Univ. are published in < < modern electronic technology > > total the 317th phase of the 6th phase in 2010 based on many features, there are following three problems in the method: 1) has ignored the text keyword in clothing information, text keyword information plays an important role for clothes classification and primary election, 2) only extract shape facility and the color characteristic in image of clothing, ignored the impact of other features on image of clothing, for example textural characteristics, 3) take the mode of layering retrieval, a feature is only used in every layer of retrieval, greatly affects result for retrieval, reduces recall ratio.
CN102254043 discloses a kind of image of clothing search method based on Semantic mapping, there is following defect in this invention: the extraction image of clothing low-level image feature 1) proposing builds garment industry knowledge base, obtains in the process of semantic information of image of clothing and exist visual signature to arrive the conversion wide gap of semantic information; 2) propose to obtain by threshold values is set the high frequency words of image of clothing in a certain feature class, but this high frequency words can not be described all clothes in this type of accurately, namely there is the not reciprocity problem of visual signature and high frequency words, accuracy rate and the recall ratio of impact retrieval.
Summary of the invention
For addressing the above problem, the invention provides a kind of image of clothing search method based on multi-modal.
Concrete technical scheme of the present invention is: a kind of image of clothing search method based on multi-modal, and specific implementation step is:
S1: build image of clothing text library;
S2: build image of clothing feature database;
S3: user inputs image of clothing to be retrieved and textual description;
S4: obtain Candidate Set Q1 by text retrieval;
S5: obtain Candidate Set Q2 by the many characteristic key of image;
S6: Candidate Set Q2 is reset and obtains result set R according to user feedback.
Further, the concrete steps of the structure image of clothing text library described in step S1 are:
S11: by web page analysis, extract the web page text information of corresponding image of clothing;
S12: by text information storage in database;
S13: text message is carried out to participle;
S14: create text index file.
Further, the concrete steps of the structure image of clothing feature database described in step S2 are:
S21: image of clothing is carried out to pretreatment operation, comprise gray scale processing and yardstick normalization process;
S22: the shape facility, color characteristic and the textural characteristics that extract respectively image;
S23: the new feature that three kinds of characteristics of image are spliced carries out cluster;
S24: obtain every corresponding cluster centre of image of clothing and describe;
S25: create image index file.
Further, image of clothing described in step S3 is the image that comprises clothes that upload this locality, and described textual description can be inputted self-defined text and acquires by choosing clothes classification or user.
Further, described in step S4, by text retrieval, obtaining Candidate Set Q1, is to utilize clothes classification that user chooses or the self-defined text of input, by text retrieval interface, obtains corresponding retrieval Candidate Set Q1.
Further, the many characteristic key of image that pass through described in step S5 obtain Candidate Set Q2, and concrete steps are:
S51: image of clothing to be retrieved is carried out to pretreatment operation, comprise gray scale processing and yardstick normalization;
S52: the shape facility, color characteristic and the textural characteristics that extract image of clothing to be retrieved;
S53: the cluster centre that obtains image to be retrieved according to the cluster centre of feature database is described;
S54: by many characteristic key of image interface, all candidate figure in atlas Q1 are screened, obtain the same or analogous Candidate Set Q2 of characteristics of image.
Further, Candidate Set Q2 is reset and obtains result set R according to user feedback described in step S6, concrete steps are:
S61: through step S5, system feedback is to the preliminary Candidate Set Q2 of user, and user selects to meet the own clothes of retrieving requirement and submits to system as feedback information from Candidate Set Q2;
S62: system acquisition user's feedback information, by data base querying, to its specific features, describe;
S63: the characteristic distance by other images in calculating and Candidate Set Q2, obtains the similarity between image of clothing;
S64: according to similarity from high to low, image in Candidate Set Q2 is rearranged, obtain net result collection R.
The invention provides the image of clothing search method based on multi-modal of a kind of comprehensive clothes text message and image information.The method is inputted image of clothing to be retrieved and textual description according to user, obtains Candidate Set, then according to user feedback, Candidate Set is reset and obtained final result for retrieval by text retrieval and the many characteristic key of image.The present invention's description to clothes by text and two kinds of mode of image of clothing, builds two-layer retrieval and rearrangement process, improves recall ratio and the precision ratio of image of clothing retrieval, has also improved recall precision simultaneously.
Accompanying drawing explanation
Fig. 1 is the costume retrieval method flow diagram based on multi-modal;
Fig. 2 is that garment feature storehouse builds schematic flow sheet;
Fig. 3 is image of clothing result for retrieval schematic diagram.
Embodiment
Below in conjunction with accompanying drawing, the present invention is described in further detail, the present embodiment is that the engineering project based on instantiation is set forth, and associated picture result for retrieval is illustrated in Fig. 3.
As shown in Figure 2, comprise 10,000 multiple image of clothing in image library, image is through pre-service, and obtaining picture size size is the gray level image of 128 pixel * 128 pixels.In the present invention, preprocessing process comprises image of clothing background Processing for removing.
As described in Figure 1, a kind of image of clothing search method based on multi-modal provided by the invention, the specific implementation step of the method is:
S1: build image of clothing text library, by web page analysis, analyze clothes shopping website on the market, extract the web page text information of corresponding image of clothing, and image of clothing is downloaded and preserved; Then text information storage is arrived to database, then text message is carried out to participle, create text index file.Described text adopts Lucene technology to set up text index, and described participle adopts JE participle technique.
S2: build image of clothing feature database, first image of clothing is carried out to pretreatment operation, extract respectively shape facility, color characteristic and the textural characteristics of image, and the splicing result of this three category feature of image of clothing is carried out to cluster, then according to cluster centre, image of clothing is described, and creates image index file.The distance that described cluster process relates between proper vector is calculated, and the present invention adopts Euclidean distance.In addition, the many features of clothes adopt K-means to carry out cluster; Characteristics of image index adopts inverted index structure.
S3: user inputs image of clothing to be retrieved and textual description, wherein image of clothing is the image that comprises clothes that upload this locality, textual description can be inputted self-defined text and acquires by choosing clothes classification or user.The form of described image of clothing is the picture format that the opencv such as BMP, JPEG, TIFF, GIF support.
S4: obtain Candidate Set Q1 by text retrieval, comprise the clothes classification or the self-defined text that utilize user to choose, by text retrieval interface, obtain corresponding retrieval set, i.e. Candidate Set Q1.Wherein, the query interface that text retrieval interface is used Lucene to provide, needs to revise returning results as comprising all satisfactory clothing information of the text field that is retrieved.
S5: obtain Candidate Set Q2 by the many characteristic key of image, first image of clothing to be retrieved is carried out to pretreatment operation, extract shape facility, color characteristic and the textural characteristics of image of clothing to be retrieved, then according to the cluster centre of feature database, obtain the cluster centre description of image to be retrieved, by many characteristic key of image interface, all candidate figure in Candidate Set Q1 are screened, obtain and the same or analogous result set of image of clothing to be retrieved, i.e. Candidate Set Q2.Wherein, all candidate figure in Candidate Set Q1 are screened, obtaining with the process of the same or analogous Candidate Set Q2 of image of clothing to be retrieved is to determine by threshold values is set, and described threshold value is recently determined by percentage.Calculate in the present embodiment front 8 cluster centres the highest with image of clothing characteristic similarity to be retrieved, according to the index file in characteristics of image storehouse, from Candidate Set Q1, filter out and the same or analogous Candidate Set Q2 of image of clothing to be retrieved.Front 8 cluster centres that image of clothing characteristic similarity described and to be retrieved is the highest are obtained by K*T, and described K is cluster centre sum, and described T is not more than 0.2 natural number, gets 0.15 in the present embodiment.
S6: Candidate Set Q2 is reset and obtains result set R according to user feedback, user selects to meet the clothes that oneself retrieval requires and submits to system as feedback information from Candidate Set Q2, system is according to user feedback, Query Database obtains feeding back the specific features of clothes and describes F, then the distance of the feature of other images in calculated characteristics F and Candidate Set Q2, obtain the similarity between image of clothing, image in Candidate Set Q2 is rearranged from high to low according to similarity, obtain net result collection R.
Wherein, between image, characteristic similarity adopts Euclidean distance to describe, and the larger similarity of distance is lower, and the less similarity of contrary distance is higher.
User feedback and the rearrangement process of acquiescence are provided in addition, in the present invention.If user carries out artificial rearrangement operation to candidate's clothing information, system passes to system and carries out rearrangement process using the clothing information of choosing for the first time according to user as feedback.This strategy can reduce user and system do not understood and caused the undesirable situation of ranking results.
In sum, the invention provides a kind of image of clothing search method based on multi-modal, according to user input text, utilize Lucene to retrieve, obtain Candidate Set Q1, and then according to similarity between image to be retrieved and other image of clothing, Q1 is screened, obtain Candidate Set Q2, finally according to user feedback, Candidate Set is reset again, obtained result set R.The present invention's description to clothes by text and two kinds of mode of image of clothing, builds two-layer retrieval and rearrangement process, improves recall ratio and the precision ratio of image of clothing retrieval, has also improved recall precision simultaneously.
Those of ordinary skill in the art will appreciate that, embodiment described here is in order to help reader understanding's principle of the present invention, should be understood to that protection scope of the present invention is not limited to such special statement and embodiment.Those of ordinary skill in the art can make various other various concrete distortion and combinations that do not depart from essence of the present invention according to these technology enlightenments disclosed by the invention, and these distortion and combination are still in protection scope of the present invention.
Claims (7)
1. the image of clothing search method based on multi-modal, is characterized in that, the concrete steps of its realization are:
S1: build image of clothing text library;
S2: build image of clothing feature database;
S3: user inputs image of clothing to be retrieved and textual description;
S4: obtain Candidate Set Q1 by text retrieval;
S5: obtain Candidate Set Q2 by the many characteristic key of image;
S6: Candidate Set Q2 is reset and obtains result set R according to user feedback.
2. the image of clothing search method based on multi-modal according to claim 1, is characterized in that, step S1 is step by step specifically:
S11: by web page analysis, extract the web page text information of corresponding image of clothing;
S12: by text information storage in database;
S13: text message is carried out to participle;
S14: create text index file.
3. the image of clothing search method based on multi-modal according to claim 2, is characterized in that, step S2 is step by step specifically:
S21: image of clothing is carried out to pretreatment operation, comprise gray scale processing and yardstick normalization process;
S22: the shape facility, color characteristic and the textural characteristics that extract respectively image;
S23: the new feature that three kinds of characteristics of image are spliced carries out cluster;
S24: obtain every corresponding cluster centre of image of clothing and describe;
S25: create image index file.
4. the image of clothing search method based on multi-modal according to claim 3, it is characterized in that, image of clothing described in step S3 is the image that comprises clothes that upload this locality, and described textual description can be inputted self-defined text and acquires by choosing clothes classification or user.
5. the image of clothing search method based on multi-modal according to claim 4, it is characterized in that, described in step S4, by text retrieval, obtain Candidate Set Q1, to utilize clothes classification that user chooses or the self-defined text of input, by text retrieval interface, obtain corresponding retrieval Candidate Set Q1.
6. the image of clothing search method based on multi-modal according to claim 5, is characterized in that, step S5 is step by step specifically:
S51: image of clothing to be retrieved is carried out to pretreatment operation, comprise gray scale processing and yardstick normalization;
S52: the shape facility, color characteristic and the textural characteristics that extract image of clothing to be retrieved;
S53: the cluster centre that obtains image to be retrieved according to the cluster centre of feature database is described;
S54: by many characteristic key of image interface, all candidate figure in atlas Q1 are screened, obtain the same or analogous Candidate Set Q2 of characteristics of image.
7. the image of clothing search method based on multi-modal according to claim 6, is characterized in that, step S6 is step by step specifically:
S61: through step S5, system feedback is to the preliminary Candidate Set Q2 of user, and user selects to meet the own clothes of retrieving requirement and submits to system as feedback information from Candidate Set Q2;
S62: system acquisition user's feedback information, by data base querying, to its specific features, describe;
S63: the characteristic distance by other images in calculating and Candidate Set Q2, obtains the similarity between image of clothing;
S64: according to similarity from high to low, image in Candidate Set Q2 is rearranged, obtain net result collection R.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410214929.4A CN103955543A (en) | 2014-05-20 | 2014-05-20 | Multimode-based clothing image retrieval method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410214929.4A CN103955543A (en) | 2014-05-20 | 2014-05-20 | Multimode-based clothing image retrieval method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103955543A true CN103955543A (en) | 2014-07-30 |
Family
ID=51332818
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410214929.4A Pending CN103955543A (en) | 2014-05-20 | 2014-05-20 | Multimode-based clothing image retrieval method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103955543A (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104331513A (en) * | 2014-11-24 | 2015-02-04 | 中国科学技术大学 | High-efficiency prediction method for image retrieval performance |
| CN104730930A (en) * | 2015-01-16 | 2015-06-24 | 小米科技有限责任公司 | Clothes sorting method and device and clothes washing method and device |
| CN105260385A (en) * | 2015-09-10 | 2016-01-20 | 上海斐讯数据通信技术有限公司 | Picture retrieval method |
| CN106250431A (en) * | 2016-07-25 | 2016-12-21 | 华南师范大学 | A kind of Color Feature Extraction Method based on classification clothing and costume retrieval system |
| CN106529606A (en) * | 2016-12-01 | 2017-03-22 | 中译语通科技（北京）有限公司 | Method of improving image recognition accuracy |
| CN107705066A (en) * | 2017-09-15 | 2018-02-16 | 广州唯品会研究院有限公司 | Information input method and electronic equipment during a kind of commodity storage |
| CN109299394A (en) * | 2018-11-26 | 2019-02-01 | Oppo广东移动通信有限公司 | Information-pushing method and Related product |
| CN111506758A (en) * | 2020-04-16 | 2020-08-07 | 腾讯科技（深圳）有限公司 | Method and device for determining article name, computer equipment and storage medium |
| CN112015923A (en) * | 2020-09-04 | 2020-12-01 | 平安科技（深圳）有限公司 | Multi-mode data retrieval method, system, terminal and storage medium |
| CN112016324A (en) * | 2020-09-03 | 2020-12-01 | 中国计量大学 | E-commerce service defect assessment method based on network comment text and picture |
| CN112148831A (en) * | 2020-11-26 | 2020-12-29 | 广州华多网络科技有限公司 | Image-text mixed retrieval method and device, storage medium and computer equipment |
| CN112347289A (en) * | 2019-08-06 | 2021-02-09 | Tcl集团股份有限公司 | Image management method and terminal |
| CN113326388A (en) * | 2021-05-20 | 2021-08-31 | 上海云从汇临人工智能科技有限公司 | Data retrieval method, system, medium and device based on inverted list |
| CN113343015A (en) * | 2021-05-31 | 2021-09-03 | 北京达佳互联信息技术有限公司 | Image query method and device, electronic equipment and computer readable storage medium |
| WO2021180109A1 (en) * | 2020-03-10 | 2021-09-16 | 华为技术有限公司 | Electronic device and search method thereof, and medium |
| CN113868442A (en) * | 2021-08-26 | 2021-12-31 | 北京中知智慧科技有限公司 | Joint retrieval method and device |
| CN115391588A (en) * | 2022-10-31 | 2022-11-25 | 阿里巴巴（中国）有限公司 | Fine adjustment method and image-text retrieval method of visual language pre-training model |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010250637A (en) * | 2009-04-17 | 2010-11-04 | Seiko Epson Corp | Image server, image retrieval system, image retrieval method, and image management method |
| CN101957825A (en) * | 2009-07-17 | 2011-01-26 | 谈玺 | Method for searching image based on image and video content in webpage |
| CN102254043A (en) * | 2011-08-17 | 2011-11-23 | 电子科技大学 | Semantic mapping-based clothing image retrieving method |
| CN102270331A (en) * | 2011-08-14 | 2011-12-07 | 黄斌 | Network shopping navigating method based on visual search |
| CN103049513A (en) * | 2012-12-14 | 2013-04-17 | 杭州淘淘搜科技有限公司 | Multi-visual-feature fusion method of commodity images of clothing, shoes and bags |
-
2014
- 2014-05-20 CN CN201410214929.4A patent/CN103955543A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010250637A (en) * | 2009-04-17 | 2010-11-04 | Seiko Epson Corp | Image server, image retrieval system, image retrieval method, and image management method |
| CN101957825A (en) * | 2009-07-17 | 2011-01-26 | 谈玺 | Method for searching image based on image and video content in webpage |
| CN102270331A (en) * | 2011-08-14 | 2011-12-07 | 黄斌 | Network shopping navigating method based on visual search |
| CN102254043A (en) * | 2011-08-17 | 2011-11-23 | 电子科技大学 | Semantic mapping-based clothing image retrieving method |
| CN103049513A (en) * | 2012-12-14 | 2013-04-17 | 杭州淘淘搜科技有限公司 | Multi-visual-feature fusion method of commodity images of clothing, shoes and bags |
Non-Patent Citations (4)
| Title |
|---|
| 刘菁华 等: "基于多媒体融合的图像检索系统的实现", 《视频技术应用与工程》 * |
| 张秀秀: "基于图像服装检索系统设计与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑（月刊）》 * |
| 王伟蔚 等: "基于DICOM文本与内容的医学图像检索系统研究", 《计算机工程与设计》 * |
| 黄崑: "综合运用基于文本与基于内容技术检索Web图像", 《情报科学》 * |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104331513A (en) * | 2014-11-24 | 2015-02-04 | 中国科学技术大学 | High-efficiency prediction method for image retrieval performance |
| CN104730930A (en) * | 2015-01-16 | 2015-06-24 | 小米科技有限责任公司 | Clothes sorting method and device and clothes washing method and device |
| CN105260385A (en) * | 2015-09-10 | 2016-01-20 | 上海斐讯数据通信技术有限公司 | Picture retrieval method |
| CN105260385B (en) * | 2015-09-10 | 2019-06-18 | 上海斐讯数据通信技术有限公司 | A kind of picture retrieval method |
| CN106250431A (en) * | 2016-07-25 | 2016-12-21 | 华南师范大学 | A kind of Color Feature Extraction Method based on classification clothing and costume retrieval system |
| CN106250431B (en) * | 2016-07-25 | 2019-03-22 | 华南师范大学 | A kind of Color Feature Extraction Method and costume retrieval system based on classification clothes |
| CN106529606A (en) * | 2016-12-01 | 2017-03-22 | 中译语通科技（北京）有限公司 | Method of improving image recognition accuracy |
| CN107705066A (en) * | 2017-09-15 | 2018-02-16 | 广州唯品会研究院有限公司 | Information input method and electronic equipment during a kind of commodity storage |
| CN107705066B (en) * | 2017-09-15 | 2022-01-07 | 唯品会(海南)电子商务有限公司 | Information input method and electronic equipment during commodity warehousing |
| CN109299394A (en) * | 2018-11-26 | 2019-02-01 | Oppo广东移动通信有限公司 | Information-pushing method and Related product |
| CN112347289A (en) * | 2019-08-06 | 2021-02-09 | Tcl集团股份有限公司 | Image management method and terminal |
| WO2021180109A1 (en) * | 2020-03-10 | 2021-09-16 | 华为技术有限公司 | Electronic device and search method thereof, and medium |
| CN111506758A (en) * | 2020-04-16 | 2020-08-07 | 腾讯科技（深圳）有限公司 | Method and device for determining article name, computer equipment and storage medium |
| CN111506758B (en) * | 2020-04-16 | 2024-05-03 | 腾讯科技（深圳）有限公司 | Method, device, computer equipment and storage medium for determining article name |
| CN112016324A (en) * | 2020-09-03 | 2020-12-01 | 中国计量大学 | E-commerce service defect assessment method based on network comment text and picture |
| CN112015923A (en) * | 2020-09-04 | 2020-12-01 | 平安科技（深圳）有限公司 | Multi-mode data retrieval method, system, terminal and storage medium |
| CN112148831B (en) * | 2020-11-26 | 2021-03-19 | 广州华多网络科技有限公司 | Image-text mixed retrieval method and device, storage medium and computer equipment |
| CN112148831A (en) * | 2020-11-26 | 2020-12-29 | 广州华多网络科技有限公司 | Image-text mixed retrieval method and device, storage medium and computer equipment |
| CN113326388A (en) * | 2021-05-20 | 2021-08-31 | 上海云从汇临人工智能科技有限公司 | Data retrieval method, system, medium and device based on inverted list |
| CN113343015A (en) * | 2021-05-31 | 2021-09-03 | 北京达佳互联信息技术有限公司 | Image query method and device, electronic equipment and computer readable storage medium |
| CN113868442A (en) * | 2021-08-26 | 2021-12-31 | 北京中知智慧科技有限公司 | Joint retrieval method and device |
| CN115391588A (en) * | 2022-10-31 | 2022-11-25 | 阿里巴巴（中国）有限公司 | Fine adjustment method and image-text retrieval method of visual language pre-training model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103955543A (en) | Multimode-based clothing image retrieval method | |
| CN104035949B (en) | Similarity data retrieval method based on locality sensitive hashing (LASH) improved algorithm | |
| KR102678975B1 (en) | Content-based search and retrieval of trademark images | |
| CN101853295B (en) | Image search method | |
| CN103886048B (en) | Cluster-based increment digital book recommendation method | |
| CN103942712A (en) | Product similarity based e-commerce recommendation system and method thereof | |
| CN104991959A (en) | Method and system for retrieving same or similar image based on content | |
| CN104834693A (en) | Depth-search-based visual image searching method and system thereof | |
| CN102254015A (en) | Image retrieval method based on visual phrases | |
| CN102339306A (en) | Sketch-based image search | |
| CN102254043A (en) | Semantic mapping-based clothing image retrieving method | |
| CN101510218A (en) | Method for implementing picture search and website server | |
| CN107766399A (en) | For the method and system and machine readable media for image is matched with content item | |
| CN105243149B (en) | A kind of semantic-based web query recommended method and system | |
| CN104965905A (en) | Web page classifying method and apparatus | |
| CN103064903A (en) | Method and device for searching images | |
| CN103631889B (en) | Image recognizing method and device | |
| CN104573130A (en) | Entity resolution method based on group calculation and entity resolution device based on group calculation | |
| Xu et al. | A semantic segmentation method with category boundary for Land Use and Land Cover (LULC) mapping of Very-High Resolution (VHR) remote sensing image | |
| CN103744887A (en) | Method and device for people search and computer equipment | |
| CN102855245A (en) | Image similarity determining method and image similarity determining equipment | |
| CN102236714A (en) | Extensible markup language (XML)-based interactive application multimedia information retrieval method | |
| CN105183737A (en) | Industrial Design instant searching system and searching method thereof | |
| CN105159898A (en) | Searching method and searching device | |
| CN110162645A (en) | Image search method, device and electronic equipment based on index |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140730 |