CN103020211A - Painting fingerprint-based artistic style retrieval system and retrieval method - Google Patents
Painting fingerprint-based artistic style retrieval system and retrieval method Download PDFInfo
- Publication number
- CN103020211A CN103020211A CN2012105249929A CN201210524992A CN103020211A CN 103020211 A CN103020211 A CN 103020211A CN 2012105249929 A CN2012105249929 A CN 2012105249929A CN 201210524992 A CN201210524992 A CN 201210524992A CN 103020211 A CN103020211 A CN 103020211A
- Authority
- CN
- China
- Prior art keywords
- database
- fingerprint
- retrieval
- purpose computer
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
The invention discloses a painting fingerprint-based artistic style retrieval system, which comprises a universal computer and a database, wherein the universal computer is connected with a scanner through a USB (Universal Serial Bus) interface and is connected with a printer through the USB interface; and the database is connected with the universal computer through 100 M Ethernet. A retrieval method of the retrieval system comprises the following specific steps of: extracting line information representing the artistic style in painting, discovering a maximum value and a minimum value in the line information by using a histogram-based statistical algorithm and storing in a painting fingerprint database; and performing painting fingerprint matching in the database to finish painting artistic style-based retrieval by using hausdorff distance algorithm as a similarity standard during retrieval. According to the system and the method, the maximum value and the minimum value in the line information are discovered and used as the painting fingerprint by using idea of computational stylistics and adopting the histogram-based statistical algorithm, the hausdorff distance algorithm is used as the similarity standard to retrieve the artistic style, and the retrieval accuracy is high.
Description
Technical field
The invention belongs to the Content-based image retrieval field, be specifically related to a kind of artistic style searching system based on the drawing fingerprint, the invention still further relates to the method that adopts said system to retrieve.
Background technology
Popularizing of Internet technology accelerated the paces of information age, impels the multimedia information technology fast development, simultaneously, also construction and the application technology of multimedia information system had higher requirement.Tool ubiquity and representational image media in the multimedia information lib, its organizational form, administrative mechanism and application technology have become the focus of current multimedia technology field research.
In order to satisfy the demand of large-scale image information management, eighties of last century has occurred manually being labeled as the image indexing system on basis, namely the semantic content of image manually marked, then adopt traditional database technique or text Information Retrieval Technology to the semantic key words of image store, index and retrieval.Rapid increase along with the view data scale, artificial mask method exposes inefficiency and the defectives such as the artificial subjectivity that marks and inconsistency gradually, has been difficult to satisfy the requirement of modern extensive image data management based on the image database system of artificial mark; The CBIR system nineties appearred in the progress along with technology such as image processing, database and artificial intelligence, and it is to gather the achievement in research of different field in the New Image administrative skill of one, and its core is based on the image similarity of content.CBIR at first extracts multiple bottom visual signature (such as the color of image, texture, shape etc.) from image, the composition characteristic vector deposits in the database.The sample retrieval that system is submitted to according to the user during retrieval returns the with it the most similar image of some in database, namely system is mapped to searching object in the feature space, so that retrieval has become the similar vector search in the feature space.
Artistic style is a very important characteristic of drawing.But because different authors' paint often has different artistic styles; and the user often shows subjectivity and changeableness to the understanding of artistic style of drawing; be different user when retrieval to the understanding of same width of cloth image with judge often can be inconsistent, even same user under different time or environment the artistic style of same width of cloth drawing is judged also can be different.Although human can understand by the perception to the visual information of drawing, the artistic style of cognitive map picture, the artistic style of how automatically understanding drawing by machine remains a very hard problem.
Summary of the invention
The purpose of this invention is to provide a kind of artistic style searching system based on the drawing fingerprint, solve existing CBIR and can not well process the restriction of the retrieval of graphics art style, realized the artistic style retrieval of any drawing, and retrieval rate is high.
Another object of the present invention provides the search method of said system.
The technical solution adopted in the present invention is, artistic style searching system based on the drawing fingerprint, comprise scanner, multi-purpose computer, printer and database, scanner is connected USB interface with multi-purpose computer and connects, printer is connected USB interface with multi-purpose computer and connects, and database is connected the 100M Ethernet and connects with multi-purpose computer;
Scanner is used for the pictorial image of needs retrieval is scanned, and scanning result is transferred in the multi-purpose computer by USB interface;
Multi-purpose computer receives the pictorial image that scans by USB interface, forms the drawing fingerprint, is transferred in the database and retrieves; Then the with it the most similar image of multi-purpose computer some that database is returned shows by display, perhaps is transferred to the printing of printer by USB;
Database is used for storage drawing fingerprint and corresponding pictorial image, and the pictorial image fingerprint of needs retrieval is mated, and returns the with it the most similar image of some.
Above-mentioned searching system search method, concrete steps are as follows:
Step 1: utilize scanner to obtain pictorial image;
Step 2: adopt multi-purpose computer to extract the line information of pictorial image: use the RADON conversion from the spatial domain of pictorial image, to extract the line information that pictorial image comprises;
Step 3: the line information according to step 2 obtains, carry out normalized;
Normalized is implemented according to following algorithm:
R=（R-min（min（R）））/（max（max（R））-min（min（R）））*225
Step 4: after normalized, adopt the statistics with histogram technology in the multi-purpose computer, obtain the statistical nature data of lines;
Step 5: the statistical nature data of the lines that obtain according to step 4, use maximum value and minimum point to set up the drawing fingerprint, and be stored in the database, the drawing fingerprint refers to all maximum value and the minimizing set in the statistical nature data;
Step 6: retrieval: extract the drawing fingerprint of width of cloth drawing, use the hausdorff distance algorithm as similarity standard, retrieve in database, database returns the with it the most similar image of some, finishes retrieval.
The invention has the beneficial effects as follows, the present invention is based on the search method of the artistic style of drawing fingerprint, by using the thought of calculating stylistics, quantize the line information of representative drawing style in the drawing, employing is based on maximum value and minimal value in the histogrammic statistic algorithm discovery line information, as the drawing fingerprint.Thereby can adopt the hausdorff distance algorithm as similarity standard, carry out the retrieval of artistic style, and retrieval rate is high.
Description of drawings
Fig. 1 is the module map of searching system of the present invention;
Fig. 2 is the process flow diagram of artistic style search method of the present invention;
Fig. 3 is width of cloth drawing, as an example;
Fig. 4 is the result of enforcement RADON shown in Figure 3 conversion;
Fig. 5 is normalization shown in Figure 4 and the result that rounds;
Fig. 6 is that enforcement shown in Figure 5 is based on the result of statistics with histogram.
Embodiment
The present invention is described in detail below in conjunction with the drawings and specific embodiments.
The present invention is based on the artistic style searching system of drawing fingerprint, referring to accompanying drawing 1, comprise scanner, multi-purpose computer, printer and database, scanner is connected USB interface with multi-purpose computer and connects, printer is connected USB interface with multi-purpose computer and connects, and database is connected the 100M Ethernet and connects with multi-purpose computer;
Scanner is used for the pictorial image of needs retrieval is scanned, and scanning result is transferred in the multi-purpose computer by USB interface;
Multi-purpose computer receives the pictorial image that scans by USB interface, forms the drawing fingerprint, is transferred in the database and retrieves; Then the with it the most similar image of multi-purpose computer some that database is returned shows by display, perhaps is transferred to printer by USB and prints;
Database is used for storage drawing fingerprint and corresponding pictorial image, and the pictorial image fingerprint of needs retrieval is mated, and returns the with it the most similar image of some.
The search method of above-mentioned searching system, concrete steps by extract the line information that represents artistic style in drawing, are used based on maximum value and minimal value in the histogrammic statistic algorithm discovery line information as shown in Figure 2, then make the drawing fingerprint database.When retrieval, adopt the hausdorff distance algorithm as similarity standard, the fingerprint matching of in database, painting, thus finish retrieval based on the graphics art style.
Concrete steps are as follows:
Step 1: utilize scanner to obtain pictorial image;
Step 2: adopt multi-purpose computer to extract the line information of pictorial image: use the RADON conversion from the spatial domain of pictorial image, to extract the line information that pictorial image comprises;
Step 3: the line information according to step 2 obtains, carry out normalized;
Normalized is implemented according to following algorithm:
R=（R-min（min（R）））/（max（max（R））-min（min（R）））*225
Step 4: after normalized, adopt the statistics with histogram technology in the multi-purpose computer, obtain the statistical nature data of lines;
Step 5: the statistical nature data of the lines that obtain according to step 4, use maximum value and minimum point to set up the drawing fingerprint, and be stored in the database, the drawing fingerprint refers to all maximum value and the minimizing set in the statistical nature data;
Step 6: retrieval: extract the drawing fingerprint of width of cloth drawing, use the hausdorff distance algorithm as similarity standard, retrieve in database, database returns the with it the most similar image of some, finishes retrieval.
The present invention is based on the search method of the artistic style of drawing fingerprint, by using the thought of calculating stylistics, the line information of representative drawing style in the quantification drawing adopts based on maximum value and minimal value in the histogrammic statistic algorithm discovery line information, as the drawing fingerprint.Thereby can adopt the hausdorff distance algorithm as similarity standard, carry out the retrieval of artistic style, and retrieval rate is high.
Embodiment
Artistic style searching system based on the drawing fingerprint, comprise scanner, multi-purpose computer, printer and database, scanner is connected USB interface with multi-purpose computer and connects, and printer is connected USB interface with multi-purpose computer and connects, and database is connected the 100M Ethernet and connects with multi-purpose computer;
Scanner is used for the pictorial image of needs retrieval is scanned, and scanning result is transferred in the multi-purpose computer by USB interface; Multi-purpose computer receives the pictorial image that scans by USB interface, forms the drawing fingerprint, is transferred in the database and retrieves; Then the with it the most similar image of multi-purpose computer some that database is returned shows by display, perhaps is transferred to printer by USB and prints; Database is used for storage drawing fingerprint and corresponding pictorial image, and the pictorial image fingerprint of needs retrieval is mated, and returns the with it the most similar image of some.
Adopt MATLAB software with concrete to a width of cloth paint fingerprint generation and be retrieved as embodiment method of the present invention be described.
Step 1: in multi-purpose computer, read in the as shown in Figure 3 pictorial image of scanner scanning, BW=imread('31.jpg'); Pictorial image is stored among the BW.
Step 2: extract line information at multi-purpose computer.Theta=0:179; [R, Xp]=radon(BW, theta); Arrange lines extract angle be 0 degree to 179 degree, the image of storing among the BW is carried out the RADON conversion, the line information of extraction is stored among the R, Xp comprises the coordinate of lines, as shown in Figure 4.
Step 3: in multi-purpose computer line information is carried out normalized, wherein normalized is implemented according to following algorithm:
R=（R-min（min（R）））/（max（max（R））-min（min（R）））*225。
Step 4: in order to improve processing speed, carry out the processing that rounds of line information; R=ceil(R); As shown in Figure 5.
Step 5: in multi-purpose computer, adopt based on histogrammic statistic algorithm, obtain the statistical nature data of lines, therefrom extract maximum value and minimal value as the drawing fingerprint, as shown in Figure 6.
Step 6: will paint fingerprint and corresponding pictorial image deposit database in, form the drawing fingerprint database, so that the retrieval of graphics art style.
Step 7: will need the drawing retrieved, and form the drawing fingerprint according to step 1-5.Then utilize the hausdorff distance algorithm as similarity standard, in database, mate.Return the with it the most similar image of some.Any 2 point set A(m) being commonly defined as of hausdorff distance B(n):
H(A，B)＝m?ax(h(A，B)，h(B，A))
Artistic style can be portrayed by quantative attribute to a certain extent.Calculating stylistics is exactly the research that reflects the drawing style by the method that quantizes.Lines are important language media that drawing consists of, and the flora development of lines has shown aesthetic trend and cultural selection of drawing, as the lines of drawing shaped core Constitution Elements, has showed philosophy theory and visual custom that drawing forms; As the lines of drawing image modelling means, the Imagery Beauty that itself comprises namely is the confirmation to the moulding of drawing image.Therefore will calculate the stylistics technology and introduce the CBIR field, with the external performance of lines as the drawing style.The search method of artistic style based on the drawing fingerprint of the present invention adopts the method that quantizes to extract the line information that represent artistic style in the drawing, adopts the statistics with histogram technology to draw the drawing finger print information, as the expression of the style of painting.The drawing fingerprint uses the hausdorff distance algorithm as similarity standard, can be used for carrying out the retrieval based on the drawing style.
Claims (2)
- Based on the drawing fingerprint the artistic style searching system, it is characterized in that, comprise scanner, multi-purpose computer, printer and database, described scanner be connected multi-purpose computer and connect by USB interface, described printer be connected multi-purpose computer and connect by USB interface, described database be connected multi-purpose computer and connect by the 100M Ethernet;Described scanner is used for the pictorial image of needs retrieval is scanned, and scanning result is transferred in the multi-purpose computer by USB interface;Described multi-purpose computer receives the pictorial image that scans by USB interface, forms the drawing fingerprint, is transferred in the database and retrieves; Then the with it the most similar image of multi-purpose computer some that database is returned shows by display, perhaps is transferred to the printing of printer by USB;Described database is used for storage drawing fingerprint and corresponding pictorial image, and the pictorial image fingerprint of needs retrieval is mated, and returns the with it the most similar image of some.
- 2. the search method of artistic style searching system based on the drawing fingerprint according to claim 1 is characterized in that concrete steps are as follows:Step 1: utilize scanner to obtain pictorial image;Step 2: adopt multi-purpose computer to extract the line information of pictorial image: use the RADON conversion from the spatial domain of pictorial image, to extract the line information that pictorial image comprises;Step 3: the line information according to step 2 obtains, carry out normalized;Normalized is implemented according to following algorithm:R=（R-min（min（R）））/（max（max（R））-min（min（R）））*225Step 4: after normalized, adopt the statistics with histogram technology in the multi-purpose computer, obtain the statistical nature data of lines;Step 5: the statistical nature data of the lines that obtain according to step 4, use maximum value and minimum point to set up the drawing fingerprint, and be stored in the database, the drawing fingerprint refers to all maximum value and the minimizing set in the statistical nature data;Step 6: retrieval: extract the drawing fingerprint of width of cloth drawing, use the hausdorff distance algorithm as similarity standard, retrieve in database, database returns the with it the most similar image of some, finishes retrieval.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210524992.9A CN103020211B (en) | 2012-12-07 | 2012-12-07 | Based on artistic style searching system and the search method of drawing fingerprint |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210524992.9A CN103020211B (en) | 2012-12-07 | 2012-12-07 | Based on artistic style searching system and the search method of drawing fingerprint |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103020211A true CN103020211A (en) | 2013-04-03 |
| CN103020211B CN103020211B (en) | 2016-01-20 |
Family
ID=47968815
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210524992.9A Expired - Fee Related CN103020211B (en) | 2012-12-07 | 2012-12-07 | Based on artistic style searching system and the search method of drawing fingerprint |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103020211B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070070254A1 (en) * | 2005-09-26 | 2007-03-29 | Cazier Robert P | Video standard determination |
| CN102778680A (en) * | 2012-06-06 | 2012-11-14 | 西安电子科技大学 | Method for imaging uniformly accelerated motion rigid group targets based on parameterization |
-
2012
- 2012-12-07 CN CN201210524992.9A patent/CN103020211B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070070254A1 (en) * | 2005-09-26 | 2007-03-29 | Cazier Robert P | Video standard determination |
| CN102778680A (en) * | 2012-06-06 | 2012-11-14 | 西安电子科技大学 | Method for imaging uniformly accelerated motion rigid group targets based on parameterization |
Non-Patent Citations (2)
| Title |
|---|
| 关晓惠: "计算机辅助的国画:分类、鉴别与系统", 《全国优秀硕士学位论文全文数据库 信息科技辑》, 15 June 2005 (2005-06-15) * |
| 梁光明等: "基于概率分布均衡技术的归一化算法及在模式识别中的应用", 《信号处理》, 30 April 2009 (2009-04-30) * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103020211B (en) | 2016-01-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hu et al. | A performance evaluation of gradient field hog descriptor for sketch based image retrieval | |
| Hu et al. | Gradient field descriptor for sketch based retrieval and localization | |
| CN109213884B (en) | Cross-modal retrieval method based on sketch retrieval three-dimensional model | |
| CN106033426B (en) | Image retrieval method based on latent semantic minimum hash | |
| CN110413816B (en) | Color Sketch Image Search | |
| Bui et al. | Scalable sketch-based image retrieval using color gradient features | |
| CN104834693A (en) | Depth-search-based visual image searching method and system thereof | |
| CN110188225B (en) | Image retrieval method based on sequencing learning and multivariate loss | |
| Datta et al. | Toward bridging the annotation-retrieval gap in image search by a generative modeling approach | |
| CN101211344A (en) | Text message ergodic rapid four-dimensional visualization method | |
| Garz et al. | Creating ground truth for historical manuscripts with document graphs and scribbling interaction | |
| Mishra et al. | Image mining in the context of content based image retrieval: a perspective | |
| Wang et al. | Constraint based region matching for image retrieval | |
| Song et al. | Analyzing scenery images by monotonic tree | |
| Zhang et al. | Context-based sketch classification | |
| Kalaiarasi et al. | Clustering of near duplicate images using bundled features | |
| CN111368126B (en) | Image retrieval-oriented generation method | |
| Shen et al. | 3D Model Search Engine Based on Lightfield Descriptors. | |
| CN112836008A (en) | Index establishing method based on decentralized storage data | |
| CN105740360B (en) | Method for identifying and searching classical titles in artwork images | |
| Fung et al. | Learning primitive and scene semantics of images for classification and retrieval | |
| CN103020211B (en) | Based on artistic style searching system and the search method of drawing fingerprint | |
| CN111178409B (en) | Image matching and recognition system based on big data matrix stability analysis | |
| CN103530656B (en) | Hidden structure learning-based image digest generation method | |
| Solli | Color Emotions in Large Scale Content Based Image Indexing |
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: 20160120 Termination date: 20161207 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |