CN101908217B - Intelligent image editing method and system suitable for tree-shaped objects - Google Patents
Intelligent image editing method and system suitable for tree-shaped objects Download PDFInfo
- Publication number
- CN101908217B CN101908217B CN2010102496323A CN201010249632A CN101908217B CN 101908217 B CN101908217 B CN 101908217B CN 2010102496323 A CN2010102496323 A CN 2010102496323A CN 201010249632 A CN201010249632 A CN 201010249632A CN 101908217 B CN101908217 B CN 101908217B
- Authority
- CN
- China
- Prior art keywords
- tree
- shaped objects
- branch
- image editing
- intelligent 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Processing Or Creating Images (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses intelligent image editing method and system suitable for tree-shaped objects. The method comprises the steps of: detecting and positioning a candidate region of a tree-shaped object; building up a tree-shaped object model, describing the geometric shape, the apparent color and the topological structure of the tree-shaped object; predefining a group of structure dictionary to describe the interlacing condition and branch condition of the tree-shaped object; determining a target branch needing to be extracted or removed and indicating the start section of the target branch; detecting and extracting the target branch in the region of the tree-shaped object by means of the tree-shaped object model and the structural dictionary from an element which corresponds to the starting section, deducing an occlusion relation of the target branch with other branches, and extracting all sub branches contained in the target branch. The invention provides a comprehensive tree-shaped object model establishment method according to the characteristics of the tree-shaped object, and realizes the precise extraction of the tree-shaped objects on the basis of deducing the branch structure and the occlusion relation.
Description
Technical field
The present invention relates to computer vision and image processing method, relate in particular to the intelligent image editing method and the system that are applicable to tree-shaped objects.
Background technology
At digital age, along with the fast development of digital image acquisition apparatus, extensively popularize and improve constantly, digital photograph works emerge in multitude, requiring of people's comparison film data source is increasingly high, and photo is handled the also enhancing day by day of demand with later stage compilation.Convenient, flexible, powerful photo-editing software can satisfy increasing digital user and professional editor personnel, Artistic Design person and other field related personnel's requirement, receives many users' favor.
Famous photo-editing software such as Adobe Photoshop are widely known, and a series of pictures editting function that it provides has greatly satisfied user's request.Microsoft also pays much attention to this field, has released digital photograph editor external member Digital Image Suite, and many image processing functions are provided, and like montage, polishing, reparation etc., also has functions such as photo management, establishment photo magic lantern film.In addition; The network user's is increasing; Also impel the online editing function to receive more and more users concern and use; The Picasa of the Snapfish of many photo resource websites such as Shutterfly, Flickr, Hewlett-Packard, Kodak's online album and Google company etc. provides the line picture editting function, and the user can more easily, personalizedly use online photo resource according to the hobby of oneself.
Picture editting's Study on Technology is a hot fields in recent years; Classical problem for example numeral is scratched figure (image matting); Image repair (image inpainting) and photo editing (Photo editing) etc. all are widely used in fields such as Flame Image Process, pattern analysis, media production, digital art, film industry, image transmission.
The stingy diagram technology of numeral is mainly studied and from former figure, is accurately extracted the foreground object of arbitrary shape or target object, and representative work is like Knockout method, Bayesian method, Poisson method, Grabcut method etc.Its basic model is to regard the color value of each image pixel as the mixing of foreground color and background color value different proportion, thereby sets up suitable model solution blending ratio value, the separation of, background preceding to realize.Yet existing these technology are devoted to how accurately to take zones such as object edge, particularly hair usually, and the edge, location is difficulty very; But these class methods generally are applicable to common 2D zone and object.For elongated, branch, staggered tree-shaped objects, also be difficult to reach extraction effect desirable, intelligence, be a very problem of difficulty such as wanting from many staggered limbs, to extract some, existing method also is difficult to address this problem.
Image repair is how to study some is damaged the zone and restores to the original state, or fills up complete being removed the zone in the image, and it is consistent that zone and peripheral region after the reparation merged, and entire image is perfection naturally.Because traditional inpainting method generally is suitable for the repairing of tiny zone (like cut, literal etc.), people had proposed the method to removal, replacement and the repairing of big regional (like whole object, a large area region) again afterwards.These reparations and the technology of filling up comprise that mainly literal wipes, and block removal, old photo and the removal of old film cut, image zoom, specific objective and remove, dazzle the eyes part to eliminate or the like.Representational work comprises PDE (PDE) model, based on the method for repairing and mending of sample (example-based), based on picture editting's technology of Poisson equation (Poisson equation) etc.Though these methods have been studied for many years and a lot of proven technique are arranged, still exist some problems as slim-lined construction being repaired the method that lacks robust, can not being judged for the situation of blocking and cause unreasonable appearance as a result etc.
In a word, for this type of dendroid object special construction, existing method is difficult to be suitable for, and needs to propose a kind of brand-new model and method, handles to the characteristics of this type objects.
Summary of the invention
The object of the present invention is to provide a kind of intelligent image editing method and system that is applicable to tree-shaped objects.
The invention discloses a kind of intelligent image editing method that is applicable to tree-shaped objects, comprise pre-treatment step, detect and tree-shaped objects candidate region, location; Tree-shaped objects modelling step is set up the tree-shaped objects model, describes tree-shaped objects geometric shape, apparent colour, topological structure; Structure dictionary definition step, one group of structure dictionary of predefine, the staggered situation and the object branch situation of description tree-shaped objects; Extraction step, confirming needs to extract the target branch of perhaps removing, and indicates the The initial segment of said target branch; Begin from the pairing primitive of said The initial segment; Utilize said tree-shaped objects model and said structure dictionary; In the candidate of said tree-shaped objects zone, detect and extract said target branch, infer said target branch and other hiding relation, extract all sons that said target branch comprised.
Above-mentioned intelligent image editing method behind the preferred said extraction step, also comprises montage and repairs step that this step adopts the layered image reparation to fill up and erased the zone, the new images after obtaining editing according to customer interaction information.
Above-mentioned intelligent image editing method in the preferred said pre-treatment step, utilizes match tracing or image segmentation to detect and tree-shaped objects candidate region, location.
Above-mentioned intelligent image editing method in the preferred said tree-shaped objects modelling step, through setting up a Markov Tree, is described geometric shape, apparent colour and the topological structure of object, as the prior distribution model of tree-shaped objects.
Above-mentioned intelligent image editing method in the preferred said structure dictionary definition step, through concluding multiple alternation sum bifurcated type, has defined the apparatus derivatorius dictionary.
Above-mentioned intelligent image editing method in the preferred said extraction step, uses Ancestral sampling method, begins from the said The initial segment of said target branch, obtains thereafter for primitive through sampling; In each step sampling process, infer structure, and be under the prerequisite of branched structure, infer the hierarchical relational between the different branches in structure when anterior branch when anterior branch.
Above-mentioned intelligent image editing method; In preferred said montage and the repairing step; Instruct down at the hierarchical relationship of inferring and estimating, to carried out background layer and other layering repairing that is retained tree-shaped objects by repairing area, recover be connected former figure in deleted the object that arborescence blocks.
On the other hand, the invention also discloses a kind of intelligent image editing system that is applicable to tree-shaped objects, comprising: pre-processing module is used for detecting and tree-shaped objects candidate region, location; The tree-shaped objects model building module is used to set up the tree-shaped objects model, describes tree-shaped objects geometric shape, apparent colour, topological structure; Structure dictionary definition module is used for one group of structure dictionary of predefine, describes the staggered situation and the object branch situation of tree-shaped objects; Extraction module is used for confirming that needs extract or the target branch of removal, and indicates the The initial segment of said target branch; Begin from the pairing primitive of said The initial segment; Utilize said tree-shaped objects model and said structure dictionary; In said tree-shaped objects candidate region, detect and extract said target branch, infer said target branch and other hiding relation, extract all sons that said target branch comprised.
Above-mentioned intelligent image editing system behind the preferred said extraction module, also is connected with montage and repairs module, and this module is used for according to customer interaction information, adopts the layered image reparation to fill up and is erased the zone, the new images after obtaining editing.
Above-mentioned intelligent image editing system in the preferred said pre-processing module, utilizes match tracing or image segmentation to detect and tree-shaped objects candidate region, location.
Geometric shape, apparent colour and the topological structure of object in the preferred said tree-shaped objects model building module, through setting up a Markov Tree, described, as the prior distribution model of tree-shaped objects by above-mentioned intelligent image editing system.
Above-mentioned intelligent image editing system in the preferred said structure dictionary definition module, through concluding multiple alternation sum bifurcated type, has defined the apparatus derivatorius dictionary.
Above-mentioned intelligent image editing system in the preferred said extraction module, uses Ancestral sampling method, begins from the said The initial segment of said target branch, obtains thereafter for primitive through sampling; In each step sampling process, infer structure, and be under the prerequisite of branched structure, infer the hierarchical relational between the different branches in structure when anterior branch when anterior branch
Above-mentioned intelligent image editing system; In preferred said montage and the repairing module; Instruct down at the hierarchical relationship of inferring and estimating, to carried out background layer and other layering repairing that is retained tree-shaped objects by repairing area, recover be connected former figure in deleted the object that arborescence blocks.
In terms of existing technologies, the invention has the advantages that:
(grown form is elongate according to these type objects characteristics in the present invention; Usually contain branch; Numerous " limbs " project on the two dimensional image plane interlaced; Difficult each other the branch) proposed comprehensive tree-shaped objects modeling method, on the basis of inferring its apparatus derivatorius, hiding relation (hierarchical relationship), realized accurate extraction tree-shaped objects; And realize down intelligent montage alternately in simple user, wipe out specific " branch " that the user hopes to erase; And the image after can editing the user carries out high-quality auto-mending.
Description of drawings
Fig. 1 a is applicable to the flow chart of steps of the intelligent image editing method embodiment of tree-shaped objects for the present invention;
Fig. 1 b is applicable to the flow chart of steps of the intelligent image editing method embodiment of tree-shaped objects for the present invention;
Fig. 2 is candidate's tree-shaped objects primitive, is staggered in together, in image, can form the structural representation of dissimilar joints.
Fig. 3 is applicable to the structural representation of the intelligent image editing system embodiment of tree-shaped objects for the present invention.
Embodiment
For make above-mentioned purpose of the present invention, feature and advantage can be more obviously understandable, below in conjunction with accompanying drawing and embodiment the present invention done further detailed explanation.
With reference to Fig. 1 a, Fig. 1 b, Fig. 1 a are applicable to the flow chart of steps of the intelligent image editing method embodiment of tree-shaped objects for the present invention; The detailed process process synoptic diagram of Fig. 1 b Fig. 1 a.Comprise the steps:
Pre-treatment step S11 detects and tree-shaped objects candidate region, location; Tree-shaped objects modelling step S12 sets up the tree-shaped objects model, describes tree-shaped objects geometric shape, apparent colour, topological structure; Structure dictionary definition step S13, one group of structure dictionary of predefine, the staggered situation and the object branch situation of description tree-shaped objects; Extraction step S14, confirming needs to extract the target branch of perhaps removing, and indicates the The initial segment of said target branch; Begin from the pairing primitive of said The initial segment; Utilize said tree-shaped objects model and said structure dictionary; In said tree-shaped objects candidate region, detect and extract said target branch, infer said target branch and other hiding relation, extract all sons that said target branch comprised.
With reference to Fig. 1 b; To given input picture, we provide two kinds of methods to obtain the initial alignment of all candidate's prospect tree-shaped objects and cut apart: (a) utilize match tracing (Matching Pursuit) method (S.Mallat and Z.Zhang, " Matching Pursuit in a Time-Frequency Dictionary " among the pre-treatment step S11; IEEE Sig.Proc.; 41,3397-415,1993.) detect all tree-shaped objects primitives (potential segment primitive); (b) utilize image Segmentation Technology; Like Lazy snapping method (Lazy Snapping.Yin Li; Jian Sun, Chi-Keung Tang, Heung-Yeung Shum; SIGGRAPH 2004.) obtain the initial segmentation in whole prospect tree-shaped objects zone, and also to be organized into the section of cutting apart basically be the tree-shaped objects primitive of unit.Obtain some candidate's tree-shaped objects zone through this step.Through setting up markov tree (Markov Tree) model, geometric shape, apparent colour and the apparatus derivatorius of object described comprehensively, among the tree-shaped objects modelling step S12 as the prior distribution model of tree-shaped objects.S13 is through concluding some kinds of typical alternation sum bifurcated types for structure dictionary definition step, has defined a series of apparatus derivatorius dictionaries, is used for tree-shaped objects and detects.Based on the apparatus derivatorius dictionary that defines among model that defines among the step S12 and the step S13; Extraction step S14 uses Ancestral sampling method (C.Bishop.Pattern Recognition and MachineLearning.Springer; 2006.); Begin from target " branch " start-up portion of user's indication, obtain its " offspring " primitive through sampling.In each step sampling process, infer the hierarchical relational between structure when anterior branch (branch or not branch) and the different branch (if existence).
If the user hopes to remove certain " branch ", then get into montage and repair step S15.The image repair technology (Layered imageinpainting) that the present invention proposes a kind of layering in this step is filled up and is erased part; According to the hierarchical relationship among the extraction step S14; Utilize respectively and wait to fill up relevant each layer information in zone and carry out the layering reparation; Make that be removed the zone coincide with background, and make originally owing to target " branch " block the arborescence that breaks off can be by reasonable connection and recovery.Indicate whether that according to customer interaction information needs remove other branch, then proceed beta pruning in this way and fill up that this process can be carried out until meeting consumers' demand repeatedly.
The tree-shaped objects modeling
Tree-shaped objects C is modeled as by some primitive segments and forms, and each primitive is the rectangle segment with certain width and length, and c representes by its central point, and C=(c1 ..., cn).Each primitive comprises geometric attribute, color attribute and structure attribute.Wherein geometric attribute comprises position coordinates, the width of this primitive; Color attribute is represented by the color histogram of this primitive region; Structure attribute comprises and the topological angle distribution at the degree of connection of this primitive number of adjacent other primitive of this primitive (promptly with), this primitive place (angle that forms with other adjacent primitive of this primitive).Set up the variation of a Markov Tree model description tree-shaped objects geometric shape, color, structure,
Wherein J representes the set of all bifurcations (promptly degree of connection is more than or equal to 3), S
jExpression c
jThe set in abutting connection with primitive,
In the formula (1) first connects to take advantage of representes the not metastasis model of branch pattern, and second company takes advantage of expression branched structure type.Branch pattern does not deteriorate to Markov chain (Markov Chain) model on one three rank, and each step metastasis model is described as the level and smooth transition of each attribute, representes with corresponding Gauss model.The branched structure type definition should the place by son branch primitive S
jWith father's branch primitive c
jBetween the variation relation of each attribute, with c
jBe benchmark; The width of each son branch primitive is modeled as being the Gauss model of expectation less than a certain amount of value of father's branch width; Color is expectation with father's branch color, and the angle distribution of this bifurcation is modeled as the higher-dimension Gauss model that blocks, and we encourage uniform angle distribution here.
Tree-shaped objects extracts
According to Bayes principle, the posterior probability of from image, extracting tree-shaped objects does
p(C|I)＝p(I|C)p(C) (2)
The present invention adopts Ancestral sampling method (C.Bishop.Pattern Recognition and Machine Learning.Springer, 2006) from candidate's tree-shaped objects primitive, to extract target " branch ".Its concrete steps are since candidate's primitive (the tree-shaped objects start-up portion of user's appointment), and progressively sampling obtains thereafter for primitive.
As previously mentioned, through pre-treatment step S11, can obtain all candidate's tree-shaped objects primitives, these primitives possibly belong to different tree-shaped objects, and they are staggered in together, in image, can form dissimilar joints (as shown in Figure 2).Can define a dictionary in view of the above, comprising the joint type of several quasi-representatives.Every type presents different apparatus derivatoriuses and hierarchical relational again.In view of the above, each step sampling is exactly the type that forms according near candidate's primitive this place, and sample in several kinds of situation that corresponding types possibly appear from dictionary: with Y type joint is example, has three kinds of situation, (a) when the anterior branch bifurcated be two sub; (b) extend (not bifurcated) when anterior branch, meet with other one " branch " (on image) and form a joint; (c) when anterior branch termination (no longer extending also no longer bifurcated).Condition of different to the explanation of image corresponding to different posterior probability (according to (2) formula).Can obtain explanation through sampling, obtain a estimation, and distinguish this branch and other branch simultaneously, infer the hiding relation between them current object structures corresponding to above-mentioned a certain situation.If do not have branch to take place, then continue the generation thereafter of sampling; If run into an apparatus derivatorius, need carry out independently above-mentioned sampling process to each son branch; Stop primitive if sample one, then this branch (or son branch) sampling process also is that leaching process finishes.
Finally obtain a tree-shaped objects
through above-mentioned sampling step and whether enough greatly to estimate its probability of happening according to (2) formula.if
would accept
thus realized the extraction of this object; Obtained the complete structure and all son branches of this object, and from other tree-shaped objects, separated it.
Layered image is repaired
In the tree-shaped objects leaching process, the hiding relation between removed tree-shaped objects and other tree-shaped objects adjacent with it is inferred out simultaneously.The tree-shaped objects that can connect and recover to be blocked in view of the above.Therefore we adopt the image repair technology of layering, make that the tree-shaped objects relation that is retained is reasonable, and the image behind the editor reaches high-quality repairing effect.
The recovery technique that the present invention adopted is based on a kind of Fast Matching method (being called the FMM method) that Telea 2004 proposes.For by being erased the zone, at first use the colouring information of background layer to carry out FMM and repair other layer conductively-closed of this zone adjacency (at this moment with); Progressively fill other layer (background layer is the bottom) then from low to high.Other layer conductively-closed during every layer of filling; Restorative procedure all adopts the FMM method.
On the other hand, the invention also discloses a kind of intelligent image editing system implementation example that is applicable to tree-shaped objects, with reference to Fig. 3, comprising: pre-processing module 31, tree-shaped objects model building module 32, structure dictionary definition module 33 and extraction module 34.Wherein, pre-processing module 31 is used for detecting and tree-shaped objects candidate region, location; Tree-shaped objects model building module 32 is used to set up the tree-shaped objects model, describes tree-shaped objects geometric shape, apparent colour, topological structure; Structure dictionary definition module 33 is used for one group of structure dictionary of predefine, describes the staggered situation and the object branch situation of tree-shaped objects; Extraction module 34 is used for confirming that needs extract or the target branch of removal, and indicates the The initial segment of said target branch; Begin from the pairing primitive of said The initial segment; Utilize said tree-shaped objects model and said structure dictionary; In said candidate's tree-shaped objects zone, detect and extract said target branch, infer said target branch and other hiding relation, extract all sons that said target branch comprises.
To sum up, advantage of the present invention and innovative point are: (1) has proposed to be applicable to the modeling method of curve-like object and apparatus derivatorius, efficiently solves that general pattern is handled and visible sensation method is inapplicable to this type objects, the problem of robust not; (2) the invention provides a kind of method of carrying out intelligent extraction and montage for mixed and disorderly staggered dendroid object, mutual simple, can on image, remove the tree-shaped objects of appointment, and not influence other other interlaced with it branch; (3) estimation of object structures and 2.1D information and deduction method; (4) a kind of high quality graphic repairing technique of the 2.1D of combination information is provided.
More than a kind of intelligent image editing method of tree-shaped objects and system of being applicable to provided by the present invention described in detail; Used specific embodiment among this paper principle of the present invention and embodiment are set forth, the explanation of above embodiment just is used for helping to understand method of the present invention and core concept thereof; Simultaneously, for one of ordinary skill in the art, according to thought of the present invention, part all can change on embodiment and range of application.In sum, this description should not be construed as limitation of the present invention.
Claims (14)
1. an intelligent image editing method that is applicable to tree-shaped objects is characterized in that, comprising:
Pre-treatment step detects and tree-shaped objects candidate region, location;
Tree-shaped objects modelling step is set up the tree-shaped objects model, describes tree-shaped objects geometric shape, apparent colour, topological structure;
Structure dictionary definition step, one group of structure dictionary of predefine, the staggered situation and the object branch situation of description tree-shaped objects;
Extraction step, confirming needs to extract the target branch of perhaps removing, and indicates the The initial segment of said target branch; Begin from the pairing primitive of said The initial segment; Utilize said tree-shaped objects model and said structure dictionary; In said tree-shaped objects candidate region, detect and extract said target branch, infer said target branch and other hiding relation, extract all sons that said target branch comprised.
2. intelligent image editing method according to claim 1 is characterized in that,
Behind the said extraction step, also comprise montage and repair step that this step adopts the layered image reparation to fill up and erased the zone, the new images after obtaining editing according to customer interaction information.
3. intelligent image editing method according to claim 2 is characterized in that,
In the said pre-treatment step, utilize match tracing or image segmentation to detect and tree-shaped objects candidate region, location.
4. intelligent image editing method according to claim 3 is characterized in that,
In the said tree-shaped objects modelling step,, geometric shape, apparent colour and the topological structure of object described, as the prior distribution model of tree-shaped objects through setting up a Markov Tree.
5. intelligent image editing method according to claim 4 is characterized in that,
In the said structure dictionary definition step,, defined the apparatus derivatorius dictionary through concluding multiple alternation sum bifurcated type.
6. intelligent image editing method according to claim 5 is characterized in that,
In the said extraction step, use Ancestral sampling method, begin, obtain thereafter for primitive through sampling from the said The initial segment of said target branch; In each step sampling process, infer structure, and be under the prerequisite of branched structure, infer the hierarchical relational between the different branches in structure when anterior branch when anterior branch.
7. intelligent image editing method according to claim 6 is characterized in that,
Said montage with repair in the step, instruct down at the hierarchical relationship of inferring and estimating, to carried out background layer and other layering repairing that is retained tree-shaped objects by repairing area, recover be connected former figure in deleted the object that arborescence blocks.
8. an intelligent image editing system that is applicable to tree-shaped objects is characterized in that, comprising:
Pre-processing module is used for detecting and tree-shaped objects candidate region, location;
The tree-shaped objects model building module is used to set up the tree-shaped objects model, describes tree-shaped objects geometric shape, apparent colour, topological structure;
Structure dictionary definition module is used for one group of structure dictionary of predefine, describes the staggered situation and the object branch situation of tree-shaped objects;
Extraction module is used for confirming that needs extract or the target branch of removal, and indicates the The initial segment of said target branch; Begin from the pairing primitive of said The initial segment; Utilize said tree-shaped objects model and said structure dictionary; In said tree-shaped objects candidate region, detect and extract said target branch, infer said target branch and other hiding relation, extract all sons that said target branch comprised.
9. intelligent image editing according to claim 8 system is characterized in that,
Behind the said extraction module, also be connected with montage and repair module, this module is used for according to customer interaction information, adopts the layered image reparation to fill up and is erased the zone, the new images after obtaining editing.
10. intelligent image editing according to claim 9 system is characterized in that,
In the said pre-processing module, utilize match tracing or image segmentation to detect and tree-shaped objects candidate region, location.
11. intelligent image editing according to claim 10 system is characterized in that,
In the said tree-shaped objects model building module,, geometric shape, apparent colour and the topological structure of object described, as the prior distribution model of tree-shaped objects through setting up a Markov Tree.
12. intelligent image editing according to claim 11 system is characterized in that,
In the said structure dictionary definition module,, defined the apparatus derivatorius dictionary through concluding multiple alternation sum bifurcated type.
13. intelligent image editing according to claim 12 system is characterized in that,
In the said extraction module, use Ancestral sampling method, begin, obtain thereafter for primitive through sampling from the said The initial segment of said target branch; In each step sampling process, infer structure, and be under the prerequisite of branched structure, infer the hierarchical relational between the different branches in structure when anterior branch when anterior branch.
14. intelligent image editing according to claim 13 system is characterized in that,
Said montage with repair in the module, instruct down at the hierarchical relationship of inferring and estimating, to carried out background layer and other layering repairing that is retained tree-shaped objects by repairing area, recover be connected former figure in deleted the object that arborescence blocks.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010102496323A CN101908217B (en) | 2010-08-10 | 2010-08-10 | Intelligent image editing method and system suitable for tree-shaped objects |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010102496323A CN101908217B (en) | 2010-08-10 | 2010-08-10 | Intelligent image editing method and system suitable for tree-shaped objects |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101908217A CN101908217A (en) | 2010-12-08 |
| CN101908217B true CN101908217B (en) | 2012-05-23 |
Family
ID=43263670
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010102496323A Expired - Fee Related CN101908217B (en) | 2010-08-10 | 2010-08-10 | Intelligent image editing method and system suitable for tree-shaped objects |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101908217B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101661628A (en) * | 2008-08-28 | 2010-03-03 | 中国科学院自动化研究所 | Method for quickly rendering and roaming plant scene |
| CN101714262A (en) * | 2009-12-10 | 2010-05-26 | 北京大学 | Method for reconstructing three-dimensional scene of single image |
-
2010
- 2010-08-10 CN CN2010102496323A patent/CN101908217B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101661628A (en) * | 2008-08-28 | 2010-03-03 | 中国科学院自动化研究所 | Method for quickly rendering and roaming plant scene |
| CN101714262A (en) * | 2009-12-10 | 2010-05-26 | 北京大学 | Method for reconstructing three-dimensional scene of single image |
Non-Patent Citations (3)
| Title |
|---|
| Jia LIU et al..Sketch-Based Branch Modeling from Images.《2009 third International Symposium on Plant Growth Modeling,Simulation,Visualization and Application》.2009,391-394. * |
| Wei Ma et al..Image-based Plant Modeling by Knowing Leaves from Their Apexes.《ICPR 2008 19th International Conference on Pattern Recognition》.2008,1-4. * |
| Yizhou Wang et al..Modeling Complex Motion by Tracking and Editing Hidden Markov Graphs.《Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition》.2004,第1卷I-856-I-863. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101908217A (en) | 2010-12-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Cignoni et al. | Meshlab: an open-source mesh processing tool. | |
| JP6395158B2 (en) | How to semantically label acquired images of a scene | |
| Xiao et al. | Image-based street-side city modeling | |
| Musialski et al. | A survey of urban reconstruction | |
| Shen et al. | Structure recovery by part assembly | |
| CN104732506A (en) | Character picture color style converting method based on face semantic analysis | |
| CN103578085B (en) | Image cavity region based on variable-block method for repairing and mending | |
| CN104299263B (en) | A kind of method that cloud scene is modeled based on single image | |
| US7191066B1 (en) | Method and apparatus for distinguishing foliage from buildings for topographical modeling | |
| CN100583158C (en) | Cartoon animation fabrication method based on video extracting and reusing | |
| CN104715451B (en) | A kind of image seamless fusion method unanimously optimized based on color and transparency | |
| CN105528614A (en) | Cartoon image layout recognition method and automatic recognition system | |
| CN111784564B (en) | Automatic image matting method and system | |
| CN104159093A (en) | Time-domain-consistent cavity region repairing method for static scene video shot in motion | |
| CN101930461A (en) | Digital image visualized management and retrieval for communication network | |
| CN103295219B (en) | Method and device for segmenting image | |
| CN103578107A (en) | Method for interactive image segmentation | |
| CN109325451A (en) | A kind of high spatial resolution arable land full-automatic extractive technique in plot based on deep learning | |
| CN101702236B (en) | Multi-target foreground segmentation method | |
| CN102800077B (en) | Bayes non-local mean image restoration method | |
| CN114202610A (en) | Oblique photography building single model texture mapping method | |
| CN101908217B (en) | Intelligent image editing method and system suitable for tree-shaped objects | |
| Song et al. | Building extraction from high resolution color imagery based on edge flow driven active contour and JSEG | |
| CN102063705B (en) | Method for synthesizing large-area non-uniform texture | |
| CN101504768B (en) | Color image fast partition method based on deformation contour model and graph cut |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120523 Termination date: 20190810 |