CN102999758A - System and method for interpreting cartoon image layout on basis of polygonal detection - Google Patents
System and method for interpreting cartoon image layout on basis of polygonal detection Download PDFInfo
- Publication number
- CN102999758A CN102999758A CN2012104569987A CN201210456998A CN102999758A CN 102999758 A CN102999758 A CN 102999758A CN 2012104569987 A CN2012104569987 A CN 2012104569987A CN 201210456998 A CN201210456998 A CN 201210456998A CN 102999758 A CN102999758 A CN 102999758A
- Authority
- CN
- China
- Prior art keywords
- straight
- cartoon image
- line segment
- polygon
- minute mirror
- 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
Images
Abstract
The invention discloses a system and a method for interpreting cartoon image layout on the basis of polygonal detection. The method is capable of completing interpretation of the cartoon image layout by automatically identifying polygons forming a storyboard frame and judging reading sequence of storyboards according to the identified geometric position among the polygons. By automatically analyzing a cartoon image inputted by a user, the storyboards in the cartoon image layout can be acquired, and the cartoon image can be read well on mobile equipment by judging reading sequence according to the geometric positions among the identified storyboards. The method for interpreting the cartoon image layout on the basis of polygonal detection is capable of automatically interpreting the cartoon image layout, solves the problem that complicated layouts cannot be processed by method of the prior art and can improve processing efficiency and user experience.
Description
Technical field
The invention belongs to document process and computer vision field, relate to a kind of cartoon image printed page understanding system and method based on polygon detecting.
Background technology
Along with e-book, popularizing and development of the digital publications such as electronic newspaper periodical published correlation technique for Contemporary Digital and proposed many application demands and challenge.Caricature is the special publication of a class, and picture group and a small amount of literal by simple humour forms usually, liked by the reader of different national boundaries all ages and classes section, and increasing reader begins to read electronics caricature document at digital terminal.At present this electron-like caricature document is usually scanned page by page by papery caricature books and obtains, directly with the image document form or image is converted into the PDF document format stores and show.
One width of cloth cartoon image can be divided into several independently subimages (part of each polygon frame and encirclement).Make the field at caricature, these independently subimage usually be called as " minute mirror ", formed by two parts: 1) role's picture exhibition, corresponding to the picture material part in the subimage; 2) role's language shows to be dialogue, i.e. character block part in the subimage.The caricature books are to set type with the page size of common paper book to design when creation at present.Caused like this a problem, when the user when the less terminal of screen (for example smart mobile phone) is read corresponding cartoon image, can't completely clearly watch whole cartoon image, need to drag back and forth up and down picture, the non-constant of reading experience, this problem have seriously restricted caricature industry digitizing process.
In order to overcome above-mentioned problem, relevant solution is suggested in document " M.Yamada et al.Comic ImageDecomposition for Reading Comics on Cellular Phones.IEEE Trans.on information and systems; 2004 " for the first time, this scheme is cartoon image to be divided in advance divide one by one mirror, shows successively each minute mirror by reading order when showing.For the second solution, some researchists attempt using computing machine automatically to finish above-mentioned task (hereinafter referred is the cartoon image printed page understanding), above-mentioned document has also proposed a kind of adversary's work point and has cut the algorithm that divides mirror to divide the mirror reading order to detect that obtains when proposing first scheme.Yet the method limitation is and must carries out the manual reading order that could automatic Identification divides mirror of cutting apart then for the caricature space of a whole page, and for present a large amount of caricature books and image files, finishing above-mentioned work by hand wastes time and energy.Document " T.Tanaka et al.Layout analysis of tree-structuredscene frames in comic images.IJCAI, 2007 " has proposed a kind of method (hereinafter referred is a minute mirror dividing method) of cartoon image being cut apart the composition mirror first.The basic assumption of the method is that caricature always can be divided into horizontal line and vertical curve and divides one by one mirror (oblique line can be divided into level and vertical two kinds according to slope), according to the order of cutting apart, caricature can be expressed as a tree-shaped structure, wherein root node just represents the view picture caricature, leaf node just represents each minute mirror, and intermediate node represents the mode cut apart.The document has proposed a kind of cut-off rule detection method of similar Hough transformation simultaneously, and T font cut-off rule has been carried out special processing.Be not difficult to find out that the limitation of this method is larger, can only process the cartoon image of specific format, for existing the complex situations of adhesion to be difficult to process between minute mirror.Document " C.Ponsard et al.Enhancing the accessibility for all ofdigital comic books.www.eminds.hci-rg.com; 2009 " has proposed a kind of minute mirror partitioning algorithm based on the watershed segmentation algorithm, main thought is to transfer cartoon image to gray level image to carry out binaryzation, then the foreground image that obtains is carried out watershed segmentation, to cutting apart the zone that obtains, remove littlely, merge overlapping greatlyr, then obtain the segmentation result of minute mirror.Yet the method can't be processed the complicated cartoon image that adhesion is arranged between minute mirror equally.
As seen, present caricature printed page analysis method most methods can only be processed the comparatively simple caricature space of a whole page, and some even necessary the dependence are by hand processed, and do not have a kind of efficient, general cartoon image printed page understanding method.
Summary of the invention
In order to realize the caricature printed page analysis of efficient general, the present invention proposes a kind of cartoon image understanding method based on polygon detecting, the method can automatically identify the polygon (great majority are quadrilateral) that forms minute mirror frame, and judge the reading order that divides between the mirror according to the geometry site between the polygon that identifies, thereby finish the caricature printed page understanding.The present invention can provide a kind of caricature printed page analysis method of efficient general, can effectively process the comparatively complicated caricature space of a whole page.
Caricature content of the present invention mainly is to scan page by page the cartoon image that obtains or the caricature bag of the picture format that provided by caricature publisher by comic books.
Technical scheme provided by the invention is as follows:
A kind of cartoon image printed page understanding system (such as Fig. 1) based on polygon detecting comprising: line segment detection module, polygon detecting module; Wherein,
Described line segment detection module is used for the straight-line segment of identification cartoon image, and records the positional information of these straight-line segments;
Described polygon detecting module according to the result that the line segment detection module detects, is spliced straight-line segment, identifies complete polygon, and divides the positional information of mirror with the polygon that identifies as correspondence; This module and line segment detection module are finished the identification and extraction of minute mirror jointly.
Described cartoon image printed page understanding system, it is characterized in that, this system also comprises the reading order detection module, this module with the result of the identification and extraction of minute mirror as input, namely with being identified minute mirror that is partitioned into as input, according to the geometry site between the polygon, minute mirror of these polygons representatives is sorted, then minute mirror after will sorting together is input in the caricature file with format information and goes together with the positional information of each minute mirror, thereby realizes the automatic Understanding of the caricature space of a whole page.
Described cartoon image printed page understanding system is characterized in that described system also comprises the polygon authentication module, and the result is undesirable if the user thinks the caricature Page Segmentation, perhaps needs to carry out result's correction, and the user uses this module to revise.
Described cartoon image printed page understanding system is characterized in that, described polygon authentication module, thus can realize automatic polygon checking by setting minute mirror that Rules Filtering identifies.
Described cartoon image printed page understanding system is characterized in that, described polygon authentication module calculates the ratio of detected length of straigh line sum and minute mirror frame girth on minute mirror frame, if greater than 50%, then by screening, otherwise does not pass through.
The present invention provides a kind of cartoon image printed page understanding method (flow process is referring to Fig. 2) based on polygon detecting simultaneously, comprises the steps:
(1) cartoon image input: the user at first obtain picture format the caricature e-file (such as: the file of the forms such as jpg, png), these files are (.zip is numbered and be collapsed into to cartoon image page by page by the caricature bag that the picture format that obtains or provided by caricature publisher is provided the papery caricature page by page, the compressed file bag of the forms such as rar), and the user with these cartoon images input system page by page;
(2) line segment detection: detect the straight-line segment in the cartoon image, the major part of these straight-line segments all is the straight-line segment that forms minute mirror frame in the cartoon image;
(3) polygon detecting (straight line splicing): after obtaining forming minute straight-line segment of mirror frame, these straight-line segment splicings are become complete minute mirror frame;
(4) divide the mirror reading order to detect: the geometry site according to the polygon straight line that identifies sorts to a minute mirror;
(5) return the printed page understanding result: with printed page understanding and analyze the polygonal geometric position obtain and the reading order between the polygon stores in the caricature file as format information, and show the user by reading terminal.
Described cartoon image printed page understanding method is characterized in that, the implementation method of step (3) is as follows:
A) first round straight-line segment polymerization: remove the redundant straight-line segment that the straight-line detection stage obtains, thereby reduce calculated amount;
B) second take turns the straight-line segment polymerization: will belong to the straight-line segment polymerization of same minute mirror frame, and form the encirclement frame of minute mirror;
C) aftertreatment: because minute mirror layout of some caricature is comparatively complicated, even generation adhesion, therefore may be broken incomplete by second a minute mirror of taking turns straight line polymerization generation, therefore, post-processing stages be carried out completion with minute mirror frame of these fragmentations by certain rule.
Effect of the present invention is: realized a kind of caricature printed page understanding method based on polygon detecting.Carry out automatic analysis by the cartoon image for user input, thereby obtain minute mirror in the caricature space of a whole page, and judge its reading order according to the geometry site between minute mirror of identification, thereby so that cartoon image can read at mobile device better.The method can the automatic Understanding caricature space of a whole page, has solved the problem that present method can not be processed the complex topology space of a whole page, can promote treatment effeciency and user and experience.
Description of drawings
Fig. 1 is system construction drawing provided by the invention;
Fig. 2 is the process flow diagram of most preferred embodiment provided by the invention;
Fig. 3 is that two straight-line segments have overlapping synoptic diagram in parallel direction in the embodiment
Fig. 4 is the synoptic diagram of last handling process
Fig. 5 is the input cartoon image of specific embodiment
Fig. 6 is the result schematic diagram of cartoon image straight-line detection
Fig. 7 is the afterwards PRELIMINARY RESULTS synoptic diagram of minute mirror identification of two-wheeled straight line polymerization
Fig. 8 is minute net result synoptic diagram of mirror identification
Fig. 9 is the net result synoptic diagram of cartoon image printed page understanding
Embodiment
The below is for the application scenarios of cartoon image as the input of caricature printed page understanding system, introduce caricature space of a whole page automatic Understanding the implementation flow process.The cartoon image here can be to be scanned page by page by the caricature books to obtain, also can be that (the caricature bag is that .zip is numbered and be collapsed into to cartoon image page by page to the cartoon image bag that provides of caricature publisher, .rar wait the compressed file bag of form, with the cartoon image that can obtain after this compressed file decompress(ion) page by page).The line segment detection module has adopted the line Segment Detection based on Eigenvalues analysis.The equipment of user U (PC, handheld device etc.) needs to install first the client software of realizing function of the present invention, and after installing, functional module of the present invention will be deployed in this client software and go.User U just can with it as input, carry out the printed page understanding work of cartoon image in this client software after obtaining cartoon image.
The overall procedure of present embodiment is: user U with cartoon image successively input system, and sends the request of caricature printed page analysis after obtaining the cartoon image file.The straight-line detection module responds should be asked, and the straight-line segment in the identification cartoon image, and these straight-line segments mainly are the straight-line segments that forms minute mirror frame.Then, the straight-line segment recognition result is sent to polygon detecting (straight line splicing) module.The polygon detecting module is carried out pre-service with the straight-line segment that is identified, and removes redundant straight-line segment and splices, and obtains minute mirror by the polygon representative.If user U selects to enable the polygon authentication module, then the polygon detecting result can show user U in Application Program Interface.User U can check and can the result of polygon detecting be screened, and send to the reading order detection module.Sort according to the relative geometry position between minute mirror, at last cartoon image is returned to user U together with the position and the reading order that divide mirror.If user U enables the polygon authentication module for selecting, the result of polygon detecting will directly send to the reading order detection module and sort, and at last cartoon image be returned to user U together with its format information (comprising minute mirror position and a reading order).
The concrete steps of present embodiment are:
(1) line segment detection
User U obtains after the cartoon image document, in its mounted realization of input client software of the present invention, and the straight-line segment in the automatic detected image document of this software meeting.Particularly, use the left end point coordinate (x of straight-line segment
1, y
1) and right endpoint coordinate (x
2, y
2) represent.Coordinate space is defined as: the upper left angle point of the cartoon image page is as true origin, and X-direction from left to right increases, and Y direction increases from top to bottom.
(2) polygon detecting (straight-line segment splicing)
Obtain after the identified straight-line segment, the polygon detecting module is spliced for these straight-line segments, becomes complete polygon, and these polygons namely represent minute mirror in the space of a whole page.Polygon detecting mainly is comprised of following steps:
At first, comprise the straight-line segment of a large amount of redundancies by the detected straight line of straight-line detection module, it is very high to account for all straight line ratios, therefore needs to remove redundant straight-line segment.The method of removing redundant straight-line segment is to use the method for straight-line segment polymerization, is about to some straight-line segments and synthesizes the straight line section.Concrete methods of realizing is in parallel direction overlapping (straight line l two to be arranged when two straight-line segments parallel to each other
1, l
2Parallel to each other, with l
1Two end points vertical projections to l
2If two end points have one at least at l
2On, it is overlapping to claim that then these two straight lines have in parallel direction, as shown in Figure 3), and between distance during less than certain threshold value, the connectedness between two straight-line segments is True; Otherwise be False.After the connectedness between all straight-line segments all is defined, use the connected component search (to be about to the straight-line segment set and to be divided into some subsets, 0 or the path of a connection of many line segment formations during any two straight-line segments in every subset can both be gathered by this), thus all straight-line segments are carried out polymerization.Concrete way is: set a variable n and its initial value is made as 1, for the straight-line segment label is set up the mark array, and initial value is set to 0.The set of traversal straight-line segment, if the reference numerals class value of certain line segment is 0, then the mark value of this line segment is set to n, and all and connective the putting in the container for the straight-line segment of True of this straight-line segment are gone, then from this container, take out the straight line section at every turn, its index value is made as n, and with all index value be 0 and and this straight-line segment is connective puts into this container for the straight-line segment of True; Until in this container without any straight-line segment, the n value is added 1, and continue to seek the reference numerals class value be 0 straight-line segment, until the set of traversal straight-line segment finishes.After the straight-line segment polymerization finished, each class was all replaced by the straight line section, had therefore reduced the number of straight-line segment by the polymerization of first round straight-line segment.
Then, carry out the polymerization second time for removing redundant straight-line segment, the purpose of this time polymerization is that the straight-line segment splicing is become rectangle.Second takes turns the mode that the connected component search is used in polymerization equally, but variation has occured in the definition of straight-line segment connectedness.For any two straight-line segments (AB, CD), wherein any one of the distance between their end points (being AC, AD, BC, BD) be less than certain less threshold value, and then the connectedness between two straight-line segments is True; Otherwise be False.After the connectedness between all straight-line segments all is defined, use the connected component search, will belong to the straight-line segment polymerization of same minute mirror encirclement frame.
At last, because the layout of some cartoon image space of a whole page is comparatively complicated, the object of minute mirror inside (sound effect word for example, to white edge etc.) can outstandingly surround frame, and therefore taking turns the straight-line segment polymerization that produces after the polymerization through second is not that complete polygon surrounds frame.The polygon detecting module infers complete polygon according to incomplete polygon encirclement frame and surrounds frame, finishes polygon identification (being the straight line splicing) work, and the result of polygon Module recognition is exported.Concrete implementation is, does certain aftertreatment according to the quantity that consists of polygonal straight-line segment, and a minute mirror that intactly identifies should comprise four straight-line segments, and the end points of each bar straight-line segment all is connected with the end points of adjacent straight-line segment.If minute mirror of identification comprises four or five straight-line segments, but the situation that exists two straight-line segment end points not to be connected with adjacent straight line segment end points then prolongs these two line segments until intersect, shown in Fig. 4 (1); For minute mirror recognition result that comprises two or three straight-line segments, these minutes the mirror recognition result may belong to same minute mirror, therefore need to carry out certain merging to these recognition results.Inspection is every a pair of by two or three recognition results that straight-line segment consists of, with (C
A, C
B) expression, find the minimum rectangle of this a pair of recognition result to surround frame Bound
AB, and their minimum rectangle are separately surrounded frame Bound
A, Bound
B, when following condition all satisfied, this a pair of recognition result can merge became a minute mirror (shown in Fig. 4 (2)):
1) Bound
ABIn the span of X-axis and Y-axis respectively less than wide and high half of source cartoon image;
2) Bound
AB, Bound
AThe area and the Bound that overlap
AThe ratio of area greater than 75%;
3) Bound
AB, Bound
BThe area and the Bound that overlap
BThe ratio of area greater than 75%.
For the remaining recognition result that comprises three straight-line segments, connect two end points that are not connected with other straight-line segments, the rectangle that forms a sealing surrounds frame, shown in Fig. 4 (3).
(3) the polygon authentication module is optional module.Under the default situations, this module is not enabled, and the recognition result that this moment, the polygon authentication module obtained will direct input as the reading order detection module.User U enables this module in the present embodiment.Then when the polygon end of identification, the polygon recognition result will be in client shows to user U.User U can check recognition result, and screens.User U preserves modification, and confirms the polygon recognition result.If use the computing machine automatic screening, concrete grammar is, because at the post-processing stages end points of the straight-line segment that goes out of linkage identification virtually, so in the process of automatic screening, calculate detected length of straigh line sum and the ratio of dividing mirror frame girth on minute mirror frame, if greater than 50%, then pass through screening, otherwise do not pass through.
(4) user U preserves and confirms after the polygon recognition result, will be as the input of reading order detection module.The reading order detection module is judged its reading order according to the geometry site between the polygon.The mirror that divides that the identification of minute mirror reading order will input sorts, and the core of ordering is how to judge any two minutes reading orders between the mirror.Specific practice is to judge two minutes horizontal and vertical overlapping situations of mirror: if 1. two minutes mirrors do not have overlappingly at vertical direction, then divide the less reading order of mirror central point Y value forward; 2. if two minutes mirrors have overlapping horizontal direction without overlapping at vertical direction, divide the larger reading order of mirror central point X value forward for Japanese caricature, after then reading order leans on for America and Europe and Hong Kong, Macao and Taiwan, continent caricature; 3. if two minutes horizontal vertical directions of mirror all have overlappingly, then two minutes mirror adjacents must have the straight line of pair of parallel, if this two straight-line segment and X-axis angle are less than 45 ° then with situation 2, otherwise with situation 1.
(5) final, client software returns to the user with caricature printed page analysis result (comprising source images, geometric position and the reading order of minute mirror that identifies).
Specific embodiment
The below for the cartoon image of width of cloth input, describes the step-by-step procedure that the present invention realizes according to above-mentioned specific implementation method.The cartoon image that at first a width of cloth is obtained by scanning (as shown in Figure 5) is as the input of system.The line segment detection module detects the straight-line segment (as shown in Figure 6) in the cartoon image; Through having obtained the PRELIMINARY RESULTS (as shown in Figure 7) that minute microscopy is surveyed after the two-wheeled cluster; Carry out aftertreatment for PRELIMINARY RESULTS afterwards, form complete minute mirror frame, and to obtain a minute microscopy survey through automatic screening be the net result (as shown in Figure 8) that cartoon image is cut apart; According to the geometry site between minute mirror, finish minute identification of mirror reading order, finally finish the cartoon image printed page understanding, net result is as shown in Figure 9.
Claims (8)
1. the cartoon image printed page understanding system based on polygon detecting is characterized in that, comprising: line segment detection module, polygon detecting module; Wherein,
Described line segment detection module is used for the straight-line segment of identification cartoon image, and records the positional information of these straight-line segments;
Described polygon detecting module according to the result that the line segment detection module detects, is spliced straight-line segment, identifies complete polygon, and divides the positional information of mirror with the polygon that identifies as correspondence; This module and line segment detection module are finished the identification and extraction of minute mirror jointly.
2. cartoon image printed page understanding as claimed in claim 1 system, it is characterized in that, described system also comprises: the reading order detection module, this module with the result of the identification and extraction of minute mirror as input, namely with being identified minute mirror that is partitioned into as input, according to the geometry site between the polygon, minute mirror of these polygon representatives is sorted, then minute mirror after will sorting together is input in the caricature file with format information and goes together with the positional information of each minute mirror, thereby realizes the automatic Understanding of the caricature space of a whole page.
3. cartoon image printed page understanding as claimed in claim 1 system, it is characterized in that described system also comprises: the polygon authentication module, the result is undesirable if the user thinks the caricature Page Segmentation, perhaps need to carry out result's correction, the user uses this module to revise.
4. cartoon image printed page understanding as claimed in claim 3 system is characterized in that, described polygon authentication module, thus can realize automatic polygon checking by setting minute mirror that Rules Filtering identifies.
5. cartoon image printed page understanding as claimed in claim 3 system is characterized in that described polygon authentication module, the detected length of straigh line sum of calculating on minute mirror frame and the ratio of dividing mirror frame girth, if greater than 50%, then by screening, otherwise do not pass through.
6. the cartoon image printed page understanding method based on polygon detecting comprises the steps:
(1) cartoon image input: the user is with cartoon image input system page by page;
(2) line segment detection: detect the straight-line segment in the cartoon image, the major part of these straight-line segments all is the straight-line segment that forms minute mirror frame in the cartoon image;
(3) polygon detecting: after obtaining forming minute straight-line segment of mirror frame, these straight-line segment splicings are become complete minute mirror frame;
(4) divide the mirror reading order to detect: the geometry site according to the polygon straight line that identifies sorts to a minute mirror;
(5) return the printed page understanding result: with printed page understanding and analyze the polygonal geometric position obtain and the reading order between the polygon stores in the caricature file as format information, and show the user by reading terminal.
7. cartoon image printed page understanding method as claimed in claim 6 is characterized in that, the implementation method of step (3) is as follows:
A) first round straight-line segment polymerization: remove the redundant straight-line segment that the straight-line detection stage obtains, thereby reduce calculated amount;
B) second take turns the straight-line segment polymerization: will belong to the straight-line segment polymerization of same minute mirror frame, and form the encirclement frame of minute mirror;
C) aftertreatment: it may be broken incomplete taking turns minute mirror that the straight line polymerization produces by second, and post-processing stages is carried out completion with minute mirror frame of these fragmentations.
8. cartoon image printed page understanding method as claimed in claim 6 is characterized in that, described cartoon image is obtained by the caricature bag decompress(ion) that the picture format that obtains or provided by caricature publisher is provided the papery caricature page by page.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012104569987A CN102999758A (en) | 2012-11-14 | 2012-11-14 | System and method for interpreting cartoon image layout on basis of polygonal detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012104569987A CN102999758A (en) | 2012-11-14 | 2012-11-14 | System and method for interpreting cartoon image layout on basis of polygonal detection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102999758A true CN102999758A (en) | 2013-03-27 |
Family
ID=47928307
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012104569987A Pending CN102999758A (en) | 2012-11-14 | 2012-11-14 | System and method for interpreting cartoon image layout on basis of polygonal detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102999758A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104504712A (en) * | 2014-12-30 | 2015-04-08 | 百度在线网络技术（北京）有限公司 | Picture processing method and device |
| CN105096299A (en) * | 2014-05-08 | 2015-11-25 | 北京大学 | Polygon detection method and device |
| CN105208183A (en) * | 2014-06-27 | 2015-12-30 | 上海玄霆娱乐信息科技有限公司 | Method enabling electronic device to display cartoons |
| CN105574524A (en) * | 2015-12-11 | 2016-05-11 | 北京大学 | Cartoon image page identification method and system based on dialogue and storyboard united identification |
| CN107209774A (en) * | 2015-02-09 | 2017-09-26 | 纳宝株式会社 | Statistics provides method and statistics provides device |
| CN111492408A (en) * | 2017-12-27 | 2020-08-04 | 株式会社OPTiM | Cartoon data display system, method, and program |
| CN111681244A (en) * | 2020-05-29 | 2020-09-18 | 山东大学 | Blade image segmentation method, system, equipment and storage medium |
| WO2023155302A1 (en) * | 2022-02-16 | 2023-08-24 | 平安科技（深圳）有限公司 | Pdf layout segmentation method and apparatus, electronic device, and storage medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1455374A (en) * | 2002-05-03 | 2003-11-12 | 三星电子株式会社 | Apparatus and method for generating 3-D cartoon |
-
2012
- 2012-11-14 CN CN2012104569987A patent/CN102999758A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1455374A (en) * | 2002-05-03 | 2003-11-12 | 三星电子株式会社 | Apparatus and method for generating 3-D cartoon |
Non-Patent Citations (5)
| Title |
|---|
| LUYUAN LI 等: "Automatic comic page segmentation based on ploygon detection", 《SPRINGER SCIENCE+BUSINESS MEDIA，LLC 2012》, 27 September 2012 (2012-09-27) * |
| MASASHI YAMADA 等: "Comic Image Decomposition for Reading Comics on Cellular Phones", 《IEICE TRANSACTIONS》, 31 January 2004 (2004-01-31) * |
| YUSUKE IN 等: "Using Fast Frame Decomposition and Sorting by Contour Tracing Mobile Phone Comic Imaging System", 《INTERNATIONAL JOURNAL OF SYSTEMS APPLICATION,ENGINEERING & DEVELOPMENT》, 28 February 2011 (2011-02-28) * |
| 刘红敏 等: "基于基元表示的多边形检测方法", 《自动化学报》, vol. 37, no. 9, 30 September 2011 (2011-09-30) * |
| 漫画SLY: "关于漫画分镜的基础知识", 《HTTP://BLOG.SINA.COM.CN/S/BLOG_59B8E1340100AHJ7.HTML》, 22 September 2008 (2008-09-22) * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105096299A (en) * | 2014-05-08 | 2015-11-25 | 北京大学 | Polygon detection method and device |
| CN105096299B (en) * | 2014-05-08 | 2019-02-26 | 北京大学 | Polygon detecting method and polygon detecting device |
| CN105208183B (en) * | 2014-06-27 | 2018-11-02 | 上海玄霆娱乐信息科技有限公司 | The method that electronic equipment shows caricature |
| CN105208183A (en) * | 2014-06-27 | 2015-12-30 | 上海玄霆娱乐信息科技有限公司 | Method enabling electronic device to display cartoons |
| CN104504712B (en) * | 2014-12-30 | 2017-08-18 | 百度在线网络技术（北京）有限公司 | Image processing method and device |
| CN104504712A (en) * | 2014-12-30 | 2015-04-08 | 百度在线网络技术（北京）有限公司 | Picture processing method and device |
| CN107209774A (en) * | 2015-02-09 | 2017-09-26 | 纳宝株式会社 | Statistics provides method and statistics provides device |
| CN105574524B (en) * | 2015-12-11 | 2018-10-19 | 北京大学 | Based on dialogue and divide the mirror cartoon image template recognition method and system that joint identifies |
| CN105574524A (en) * | 2015-12-11 | 2016-05-11 | 北京大学 | Cartoon image page identification method and system based on dialogue and storyboard united identification |
| CN111492408A (en) * | 2017-12-27 | 2020-08-04 | 株式会社OPTiM | Cartoon data display system, method, and program |
| CN111681244A (en) * | 2020-05-29 | 2020-09-18 | 山东大学 | Blade image segmentation method, system, equipment and storage medium |
| CN111681244B (en) * | 2020-05-29 | 2022-06-21 | 山东大学 | Blade image segmentation method, system, equipment and storage medium |
| WO2023155302A1 (en) * | 2022-02-16 | 2023-08-24 | 平安科技（深圳）有限公司 | Pdf layout segmentation method and apparatus, electronic device, and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102999758A (en) | System and method for interpreting cartoon image layout on basis of polygonal detection | |
| CN108537146B (en) | Print form and handwriting mixed text line extraction system | |
| CN105528614B (en) | A kind of recognition methods of the cartoon image space of a whole page and automatic recognition system | |
| CN105574524B (en) | Based on dialogue and divide the mirror cartoon image template recognition method and system that joint identifies | |
| US10191889B2 (en) | Systems, apparatuses and methods for generating a user interface by performing computer vision and optical character recognition on a graphical representation | |
| CN109002824B (en) | OpenCV-based building drawing label information detection method | |
| US8000529B2 (en) | System and method for creating an editable template from a document image | |
| Chen et al. | Hybrid page segmentation with efficient whitespace rectangles extraction and grouping | |
| US9965695B1 (en) | Document image binarization method based on content type separation | |
| CN109685052A (en) | Method for processing text images, device, electronic equipment and computer-readable medium | |
| US8693790B2 (en) | Form template definition method and form template definition apparatus | |
| CN110502985B (en) | Form identification method and device and form identification equipment | |
| JP5854802B2 (en) | Image processing apparatus, image processing method, and computer program | |
| CN101676930A (en) | Method and device for recognizing table cells in scanned image | |
| US11042734B2 (en) | Electronic document segmentation using deep learning | |
| CN106203454A (en) | The method and device that certificate format is analyzed | |
| US10713515B2 (en) | Using multiple cameras to perform optical character recognition | |
| EP2884425B1 (en) | Method and system of extracting structured data from a document | |
| CN113158895A (en) | Bill identification method and device, electronic equipment and storage medium | |
| CN115546809A (en) | Table structure identification method based on cell constraint and application thereof | |
| WO2020071558A1 (en) | Business form layout analysis device, and analysis program and analysis method therefor | |
| Suleyman et al. | An adaptive threshold algorithm for offline Uyghur handwritten text line segmentation | |
| EP1439486B1 (en) | Segmenting an image via a graph | |
| CN116052193B (en) | RPA interface dynamic form picking and matching method and system | |
| Yu et al. | An effective method for figures and tables detection in academic literature |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20130327 |