CN104252715A - Single line image-based three-dimensional reconstruction method - Google Patents
Single line image-based three-dimensional reconstruction method Download PDFInfo
- Publication number
- CN104252715A CN104252715A CN201410450000.1A CN201410450000A CN104252715A CN 104252715 A CN104252715 A CN 104252715A CN 201410450000 A CN201410450000 A CN 201410450000A CN 104252715 A CN104252715 A CN 104252715A
- Authority
- CN
- China
- Prior art keywords
- string diagram
- model
- dimensional reconstruction
- dimensional
- line
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a single line image-based three-dimensional reconstruction method. The method comprises the following steps: firstly performing vectorization processing on an input line image, and converting the line image into a two-dimensional vector line plot; then, matching the two-dimensional vector line plot and a preset three-dimensional model base by applying a subgraph isomorphism method, wherein the model matched with the two-dimensional vector line plot is called a candidate model; finally, performing minimized solving on the coordinate difference function of the line plot and the candidate model so as to select the optimal model and obtain a reconstruction result. According to the single line image-based three-dimensional reconstruction method disckosed by the invention, three-dimensional reconstruction can be effectively performed on the single line image, so that the reading experiences of related documents are greatly improved.
Description
Technical field
The invention belongs to Image processing and compute machine visual field, relate to a kind of three-dimensional reconstruction method based on single width string diagram picture.
Background technology
A large amount of solid geometry figures is there is in electronic document miscellaneous.These documents comprise teaching material, examination question, demonstration original text etc.But these solid geometry figure overwhelming majority store in a document in the mode of two-dimentional lines image, and directly cannot present the 3-D solid structure of object, cause inconvenience to the reading of reader.Particularly in today that mobile reading equipment, three-dimensional display apparatus are more and more ripe, lag behind advanced display technique with the solid geometry figure of two-dimentional lines image mode storage and display, have impact on the reading experience of reader.If can be 3 D stereo as restoration and reconstruction by these string diagrams, will greatly improve the reading experience of reader.
" string diagram " is a kind of artificial X-Y scheme, and it can describe the structure of object in clear and intuitive mode.Be different from general natural image, the structure of string diagram is often made up of point, line, has geometry clearly, is made up of parallel projection, do not have see-through feature, does not have the texture information in general nature image.The common string diagram of people comprises hand-drawing graphics, Graphing of Engineering, and CAD charts, teaching material illustration, etc.String diagram is actually the connected graph (can be divided into connected subgraph for unconnected graph to process respectively) that contains summit and limit.The algorithm carrying out three-dimensional reconstruction to string diagram generally needs a pre-service to be summit two-dimensional coordinate and fillet set Graphic Exchanging.Fillet between summit can be straight-line segment, also can be curve.
In past 20 years, there is a large amount of research based on the three-dimensional reconstruction method of single width string diagram.The method that these researchs adopt, applicable object, application scenarios are all not quite similar.By the algorithm classification that method for reconstructing adopts, have based on Optimality Criteria, based on geometry hypothetical deduction, based on methods such as divide-and-conquer strategies.Rule-based is that early stage most of method adopted, and 3 D stereo is defined as geometric object by them, and uses restraint to it by some geometrical rule, to seek best result in solution room.Conventional rule has plane rule (document " An optimization-based approach to the interpretation of single line drawings as3D wire frames.International Journal of Computer Vision, 1992. "), MSDA rule (document " Emulating the human interpretation of line-drawings as three-dimensional objects.International Journal of Computer Vision, 1991. "), MSDSM rule (document " 3D Object Recovery from 2D Images:ANew Approach.SPIE Proc.Robotics and Computer Vision, 1996. "), MEAD rule (document " 3-D interpretation of single line drawings based on entropy minimization principle.Computer Vision and Pattern Recognition, 2001. ") etc.Solving-optimizing function is generally the final step of these class methods; First method based on geometry hypothesis supposes that the 3 D stereo handled by it follows certain requirement, as contained cube angle point (document " 3D reconstruction of polyhedral objects from single parallel projections using cubic corner.Computer-Aided Design; 2011. "), symmetry of having living space (document " Inferring mirror symmetric 3D shapes from sketches.Computer-Aided Design, 2012. ") etc.On the basis of hypothesis, they are inferred the point of whole object and limit coordinate, finally obtain stereo reconstruction result.Not necessarily comprise Optimization Solution step in this method, general computation complexity is relatively low, but its assumed condition is usually harsher, is only applicable to the figure that some are special; in recent years certain methods (the document " A divide-and-conquer approach to 3D object reconstruction from line drawings.IEEE 11th International Conference on Computer Vision, 2007. " occurred, document " Decomposition of complex line drawings with hidden lines for 3Dplanar-faced manifold object reconstruction.IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011. ", document " Object cut:Complex 3d object reconstruction through line drawing separation.IEEE Conference on Computer Vision and Pattern Recognition, 2010. ") have employed the three-dimensional reconstruction problem that the strategy of dividing and ruling solves some complexity, the figure of complexity is first divided into simple essential part as far as possible by them, then apply some rules to be optimized and solve, finally the result that various piece is tried to achieve is combined and obtain the result of whole reconstruction.
Visible, the current three-dimensional reconstruction method to single width string diagram mostly can only the string diagram of processing vector, and cannot process image.Even if string diagram picture can be converted into by certain pre-service the string diagram of vector, current method also highly relies on integrality and the correctness of string diagram, some even must rely on manual process and extract string diagram, and does not have a kind of three-dimensional reconstruction method that efficiently, directly can process image, that have fault-tolerant ability.
Summary of the invention
In order to realize carrying out efficient three-dimensional reconstruction to single width string diagram picture, the present invention proposes a kind of three-dimensional reconstruction method based on single width string diagram picture.First the string diagram picture of input is carried out vectorized process by the method, be converted into two-dimensional vector string diagram, two-dimensional vector string diagram mates with the 3 d model library preset by the method then applying " subgraph match ", the model matched is called candidate family, finally the coordinate difference of string diagram and candidate family is carried out minimizing solving apart from function, to select optimum model and to draw reconstructed results.The present invention can carry out three-dimensional reconstruction to single width string diagram picture effectively.
Handling object of the present invention is the geometry lines image extracted from the electronic document of PDF or other form, or by mobile phone and the shooting of other camera installation, or by geometry lines image that scanner scanning papery teaching material obtains.
Technical scheme provided by the invention is as follows:
Based on a three-dimensional reconstruction method for single width string diagram picture, it is characterized in that, comprise the steps:
1) from input picture, vector string diagram is extracted;
2) from 3 d model library for vector string diagram chooses some candidate families;
3) fixed by the apex coordinate of vector string diagram, rotation, Pan and Zoom candidate family are to mate the two-dimentional apex coordinate of vector string diagram in three dimensions, make the variance of the coordinate on vector string diagram and each summit of candidate family reach minimum value; Then from several candidate families, the model that the variance of coordinate is minimum is selected, as the result of three-dimensional reconstruction.
Preferred:
Described three-dimensional reconstruction method, is characterized in that, step 1) implementation method be:
1.1) input picture is carried out binary conversion treatment and connected component's search;
1.2) solid line and dotted line is divided into by lines to be extracted in image to carry out the extraction of straight line;
1.3) according to the intersection point of extracted straight line, straight line is cut into line segment;
1.4) filter out unnecessary lines, obtain vector string diagram.
Described three-dimensional reconstruction method, is characterized in that, step 2) described in 3 d model library in, preserve three-dimensional model with parameterized form: three-dimensional model is the string diagram in a three dimensions, and its apex coordinate is by one group of state modulator.
Described three-dimensional reconstruction method, is characterized in that, step 2) in, use the method for Subgraph Isomorphism to choose some candidate families for vector string diagram.
Described three-dimensional reconstruction method, is characterized in that, step 2) in, need twice coupling be carried out, first time using vector string diagram as large figure, using three-dimensional model as little figure; Second time using three-dimensional model as large figure, using vector string diagram as little figure.
Described three-dimensional reconstruction method, is characterized in that, step 1) implementation method be:
1.1) carry out k-means cluster to the connected component of input picture, the encirclement frame size of employing connected component and elemental area, as its cluster attribute, are divided three classes: main frame, dotted line point, descriptive text;
1.2) use Hough transform to carry out solid line extraction to main frame connected component, use RANSAC method to carry out dotted line extraction;
1.3) remove unsettled line, stop the unnecessary lines such as line, diagonal line, generate vector string diagram.
Described three-dimensional reconstruction method, is characterized in that, step 3) implementation method be:
3.1) fixed by vector string diagram apex coordinate, rotation, Pan and Zoom candidate family are to mate the two-dimentional apex coordinate of vector string diagram in three dimensions;
3.2) calculated candidate model is through the coordinate of parallel projection to vector lines plan, and obtains the variance of model projection coordinate and string diagram apex coordinate, forms objective function and be optimized it to solve;
3.3) choose make objective function minimum candidate family as optimization model.
Described three-dimensional reconstruction method, is characterized in that, further comprising the steps of:
4) result of three-dimensional reconstruction is exported.
Effect of the present invention is: achieve a kind of three-dimensional reconstruction method based on single width string diagram picture.By extracting two-dimensional vector string diagram to input picture, and it being mated with the 3 d model library preset, selecting out candidate family, then obtaining the result of three-dimensional reconstruction by the rotation to candidate family, Pan and Zoom.The method can solve current method and directly can not process image, to shortcomings such as input string diagram integrity demands are high, can promote treatment effeciency and enrich the reading experience of mobile device user.
Accompanying drawing explanation
Fig. 1 is flow process frame diagram of the present invention;
Fig. 2 is the method flow diagram extracting vector string diagram from image of the present invention;
Fig. 3 is the schematic diagram extracting vector string diagram from image.A () input picture (b) extracts solid line (d) to connected component's cluster (c) and extracts dotted line (e) resultant vector string diagram;
Fig. 4 is default 3 d model library example (a) rectangular parallelepiped (b) rectangular pyramid (c) triangular prism (d) three terrace with edge;
Fig. 5 is that method schematic diagram (a) dotted line point (b) (c) extracting dotted line extracts straight line (d) outlier;
Fig. 6 is the schematic diagram that unnecessary lines filter;
Fig. 7 is the input file and picture of specific embodiment;
Fig. 8 is the result schematic diagram of vector extraction string diagram;
Fig. 9 is the schematic diagram of Subgraph Isomorphism matching candidate model;
Figure 10 is three-dimensional reconstruction Output rusults schematic diagram;
Embodiment
Below for using the application scenarios of file and picture as the input of three-dimensional reconstruction system, introduce specific embodiment of the invention flow process.Here file and picture is the geometry lines image extracted from the electronic document of PDF or other form, or by mobile phone and the shooting of other camera installation, or by geometry lines image that scanner scanning papery teaching material obtains.The equipment (PC, handheld device etc.) of user U needs first to install the executive software realizing function of the present invention, after installing, user U can open PDF document in software, located and cut-away view picture by mouse or gesture, automatically perform three-dimensional reconstruction work in software and Output rusults.User U can watch and rotate the three-dimensional picture exported on screen.
Specific embodiment of the invention step is (ginseng Fig. 1):
(1) vector string diagram is extracted
The flow process of method as shown in Figure 2.For the solid line in image, the maximum component in image is carried out Hough transform process; For the dotted line in image, by K-means cluster, connected component less in image is screened, then carry out extraction straight line by RANSAC method.Then calculate extract the intersection point of straight line, be line segment from point of intersection straight line cutting, and merge adjoining nodes.Eventually pass filtration treatment and remove unnecessary lines, generate final line segment and vector string diagram, Fig. 3 shows file and picture example successively, the extraction result of solid line, dotted line, and the vector string diagram generated.Be below concrete implementation method:
1.1) connected component's cluster.As shown in Fig. 3 (b), usually there is the connected component of three types in file and picture: main frame, dotted line point, and descriptive text.K-means method is used to carry out cluster to the connected component in image.If k=3, and take the encirclement frame size of connected component and elemental area as its cluster attribute.After one takes turns cluster, successfully connected component is divided into three classes, the class that wherein area is maximum is chosen as main frame, and the similar and minimum class of area is dotted line point, other be descriptive text.
1.2) solid line extracts.Have employed Hough transform and on the Canny edge of main frame, carry out lines detection based on the method for edge line segment.Because there are two Canny edges on the limit of on image, the straight line of a therefore limit extraction has two and has almost identical slope and very near distance.According to their slope and distance relation, these straight lines are merged into one.
1.3) dotted line extracts.As shown in Figure 5, RANSAC method is adopted to carry out the extraction of dotted line.First, all connected components being divided into dotted line point are reduced into its central point, then can determine straight line at every 2.Find in these straight lines and comprise the maximum straight line of interior point (distance from straight line is no more than an enough little scope), take out this straight line and point is rejected in it is comprised.Repeat above process until can not find to comprise in 3 and put above straight line.Remaining point becomes outlier.
1.4) unnecessary lines filter.For some unnecessary lines (boost line, lines that mistake extracts etc.), need they removings with the success ratio improving Model Matching.The line style of type of concrete removal is as follows:
A) unsettled line: in the vector string diagram extracted, if the degree of a lines end points is 1, then it is called unsettled line.If the lines 6-7 in Fig. 6 (a) is exactly a unsettled line, the degree of its end points 7 is 1.The vector string diagram of coordinate axis and imperfect extraction is more common in by other typical unsettled lines.Unsettled line is needed them to remove.
B) stop line: if the end points of lines is not just in time the center section of another lines (being two ends), then it is stop line (Fig. 6 (b) CE, CF, a C
1e,C
1f).Boost line in many file and pictures is all stop line.These lines also need to be removed.
C) diagonal line: another kind of unnecessary lines are diagonal line of the parallelogram in figure, as A in Fig. 6 (c)
1b, BC
1, A
1c
1.What these diagonal line destroyed originally object opens up benefit structure, therefore needs to be removed.
(2) three-dimensional model coupling
Based on the analysis to representative document image, the present invention establishes following 3 d model library: 3 d model library is made up of several three-dimensional models, and model is the string diagram in a three dimensions, and its apex coordinate is by one group of state modulator.Model in 3 d model library is all the typical solid figure in file and picture, as rectangular parallelepiped, and rectangular pyramid, three terrace with edges etc., as shown in Figure 4.For Fig. 4 (a), a rectangular parallelepiped model has three parameters: a={x, y, z}, represents that all apex coordinates of this model are with parameter matrix V
Wherein the three-dimensional coordinate on a summit is shown in each list of V, and every a line represents a dimension, and a is the controling parameters vector of model, and it controls the geometric attributes such as the length of model.Anyly comprise N
mthe model m on individual summit has 3 × N
mparameter matrix V
m.
After extraction vector string diagram completes, use VF-2 Subgraph Isomorphism algorithm in model bank, find out one group of candidate family." Subgraph Isomorphism " be a kind of in a large figure find with a little figure open up the technology of mending the consistent subgraph of structure.
The present invention needs to carry out twice Matching Model at every turn: " forward coupling " and " reverse mate ".Forward coupling refers to using vector string diagram as large figure, using three-dimensional model as little figure, finds the mode of the subgraph of isomorphism in vector string diagram.Reverse coupling is then just in time contrary, and it is using three-dimensional model as large figure, and using vector string diagram as little figure, finds isomorphism subgraph in the three-dimensional model.
Forward matching way can only be used for the vector string diagram of complete extraction.Due to the limitation of vector string diagram extraction algorithm, for the vector string diagram (as Fig. 6 (a)) not exclusively extracted, forward coupling will be failed.Therefore, use reverse matching process to solve this problem: three-dimensional model, as large figure, finds vector lines graph isomorhpism subgraph wherein.Therefore, in reverse coupling, even if vector string diagram is imperfect, also successfully candidate family can be found out.
(3) three-dimensional reconstruction
Based on the result of three-dimensional model coupling, adopt and carry out three-dimensional reconstruction with the following method: fixed by vector string diagram apex coordinate, rotation, Pan and Zoom model are to mate the two-dimentional apex coordinate of vector string diagram in three dimensions.Concrete grammar is: define a three-dimensional rotation matrix R, a translation matrix t, and the controling parameters of model is a, if each apex coordinate of institute's Matching Model is V
i, (convergent-divergent of apex coordinate is controlled by a), the apex coordinate of vector string diagram is x
i, then the coordinate of three-dimensional model after over-rotation, translation is
(RV
i+t) (1)
So it with the coordinate difference on each summit of former vector string diagram is
(RV
i+t)-x
i (2)
Target finds such
make the quadratic sum of the coordinate difference on all summits minimum, namely
The method of document " A feasible method for optimization with orthogonality constraints.Technical report, Rice University, 2010 " is used to solve relevant optimization problem.Each candidate family before selected can carry out the parameter optimization of formula (3), and be optimized result
the minimum model of target function value f is wherein made to be finalized as whole modeling type.Final reconstructed results is according to optimizing the parameter obtained by whole modeling type
carry out the rotation in three dimensions, Pan and Zoom obtains
Wherein S is the three-dimensional coordinate on each summit of Output rusults,
for rotation matrix, V be by
element form parameter matrix (
the convergent-divergent of control vertex coordinate), T serves as reasons
the translation matrix formed.
Specific embodiment
Below according to above-mentioned specific implementation method, for the file and picture of a width input, the process that the present invention realizes is described.
First from the electronic document of PDF or other form, a width geometry lines image is intercepted, or by mobile phone and the shooting of other camera installation, or by scanner to papery teaching material scanning acquisition one width geometry lines image, then using the input of the image (as shown in Figure 7) of acquisition as system.Extract the result of vector string diagram as shown in Figure 8.Then use Subgraph Isomorphism to select the candidate family (as shown in Figure 9) of coupling, and only retain the matching result of clique.Eventually through rotation, translation candidate family, and the result choosing optimization object function minimum completes three-dimensional reconstruction.Final Output rusults as shown in Figure 10.
Claims (8)
1. based on a three-dimensional reconstruction method for single width string diagram picture, it is characterized in that, comprise the steps:
1) from input picture, vector string diagram is extracted;
2) from 3 d model library for vector string diagram chooses some candidate families;
3) fixed by the apex coordinate of vector string diagram, rotation, Pan and Zoom candidate family are to mate the two-dimentional apex coordinate of vector string diagram in three dimensions, make the variance of the coordinate on vector string diagram and each summit of candidate family reach minimum value; Then from several candidate families, the model that the variance of coordinate is minimum is selected, as the result of three-dimensional reconstruction.
2. three-dimensional reconstruction method as claimed in claim 1, is characterized in that, step 1) implementation method be:
1.1) input picture is carried out binary conversion treatment and connected component's search;
1.2) solid line and dotted line is divided into by lines to be extracted in image to carry out the extraction of straight line;
1.3) according to the intersection point of extracted straight line, straight line is cut into line segment;
1.4) filter out unnecessary lines, obtain vector string diagram.
3. three-dimensional reconstruction method as claimed in claim 1, it is characterized in that, step 2) described in 3 d model library in, preserve three-dimensional model with parameterized form: three-dimensional model is the string diagram in a three dimensions, and its apex coordinate is by one group of state modulator.
4. three-dimensional reconstruction method as claimed in claim 3, is characterized in that, step 2) in, use the method for Subgraph Isomorphism to choose some candidate families for vector string diagram.
5. three-dimensional reconstruction method as claimed in claim 4, is characterized in that, step 2) in, need twice coupling be carried out, first time using vector string diagram as large figure, using three-dimensional model as little figure; Second time using three-dimensional model as large figure, using vector string diagram as little figure.
6. three-dimensional reconstruction method as claimed in claim 1, is characterized in that, step 1) implementation method be:
1.1) carry out k-means cluster to the connected component of input picture, the encirclement frame size of employing connected component and elemental area, as its cluster attribute, are divided three classes: main frame, dotted line point, descriptive text;
1.2) use Hough transform to carry out solid line extraction to main frame connected component, use RANSAC method to carry out dotted line extraction;
1.3) remove the unnecessary lines comprising unsettled line, stop line, diagonal line, generate vector string diagram.
7. three-dimensional reconstruction method as claimed in claim 1, is characterized in that, step 3) implementation method be:
3.1) fixed by vector string diagram apex coordinate, rotation, Pan and Zoom candidate family are to mate the two-dimentional apex coordinate of vector string diagram in three dimensions;
3.2) calculated candidate model is through the coordinate of parallel projection to vector lines plan, and obtains the variance of model projection coordinate and string diagram apex coordinate, forms objective function and be optimized it to solve;
3.3) choose make objective function minimum candidate family as optimization model.
8. three-dimensional reconstruction method as claimed in claim 1, is characterized in that, further comprising the steps of:
4) result of three-dimensional reconstruction is exported.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410450000.1A CN104252715B (en) | 2014-09-05 | 2014-09-05 | Single line image-based three-dimensional reconstruction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410450000.1A CN104252715B (en) | 2014-09-05 | 2014-09-05 | Single line image-based three-dimensional reconstruction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104252715A true CN104252715A (en) | 2014-12-31 |
| CN104252715B CN104252715B (en) | 2017-05-03 |
Family
ID=52187581
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410450000.1A Active CN104252715B (en) | 2014-09-05 | 2014-09-05 | Single line image-based three-dimensional reconstruction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104252715B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106004140A (en) * | 2016-05-19 | 2016-10-12 | 清华大学 | Method for displaying 3D animation in single image |
| CN107615279A (en) * | 2015-04-02 | 2018-01-19 | 海德龙斯有限公司 | Virtual three-dimensional model generation based on virtual hexahedron model |
| CN107871333A (en) * | 2017-11-30 | 2018-04-03 | 上海联影医疗科技有限公司 | Line drawing drawing method and device, computer equipment and computer-readable storage medium |
| CN110414402A (en) * | 2019-07-22 | 2019-11-05 | 北京达佳互联信息技术有限公司 | A kind of gesture data mask method, device, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101937579A (en) * | 2010-09-20 | 2011-01-05 | 南京大学 | Method for creating three-dimensional surface model by using perspective sketch |
| CN102880868A (en) * | 2012-08-06 | 2013-01-16 | 上海中和软件有限公司 | Engineering drawing vector conversion and primitive semantic extraction method |
| CN103390088A (en) * | 2013-07-31 | 2013-11-13 | 浙江大学 | Full-automatic three-dimensional conversion method aiming at grating architectural plan |
-
2014
- 2014-09-05 CN CN201410450000.1A patent/CN104252715B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101937579A (en) * | 2010-09-20 | 2011-01-05 | 南京大学 | Method for creating three-dimensional surface model by using perspective sketch |
| CN102880868A (en) * | 2012-08-06 | 2013-01-16 | 上海中和软件有限公司 | Engineering drawing vector conversion and primitive semantic extraction method |
| CN103390088A (en) * | 2013-07-31 | 2013-11-13 | 浙江大学 | Full-automatic three-dimensional conversion method aiming at grating architectural plan |
Non-Patent Citations (6)
| Title |
|---|
| CHANGQING ZOU ET AL: "Separation of line Drawings Based on split faces for 3D objects reconstruction", 《2014 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION》 * |
| XUE TIANFAN ET AL: "Example-Based 3D object reconstruction from line Drawings", 《2012 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION(CVPR)》 * |
| 刘滢滢: "栅格图像矢量化研究与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 孙玉红 等: "线条画矢量化研究", 《电子技术》 * |
| 裘尧波: "工程图纸扫描图像矢量化方法的设计与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 郑金鑫: "基于单幅线条图的三维立体重建方法综述", 《计算机科学》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107615279A (en) * | 2015-04-02 | 2018-01-19 | 海德龙斯有限公司 | Virtual three-dimensional model generation based on virtual hexahedron model |
| CN107615279B (en) * | 2015-04-02 | 2020-11-03 | 李明学 | Virtual three-dimensional model generation based on virtual hexahedron model |
| CN106004140A (en) * | 2016-05-19 | 2016-10-12 | 清华大学 | Method for displaying 3D animation in single image |
| CN107871333A (en) * | 2017-11-30 | 2018-04-03 | 上海联影医疗科技有限公司 | Line drawing drawing method and device, computer equipment and computer-readable storage medium |
| CN110414402A (en) * | 2019-07-22 | 2019-11-05 | 北京达佳互联信息技术有限公司 | A kind of gesture data mask method, device, electronic equipment and storage medium |
| CN110414402B (en) * | 2019-07-22 | 2022-03-25 | 北京达佳互联信息技术有限公司 | Gesture data labeling method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104252715B (en) | 2017-05-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110443892B (en) | Three-dimensional grid model generation method and device based on single image | |
| Zheng et al. | Non-local scan consolidation for 3D urban scenes | |
| US20210027532A1 (en) | Primitive-based 3d building modeling, sensor simulation, and estimation | |
| Qiu et al. | Pipe-run extraction and reconstruction from point clouds | |
| CN109902583B (en) | Skeleton gesture recognition method based on bidirectional independent circulation neural network | |
| EP3474185B1 (en) | Classification of 2d images according to types of 3d arrangement | |
| Jo et al. | Data augmentation using synthesized images for object detection | |
| CN104252715B (en) | Single line image-based three-dimensional reconstruction method | |
| Gao et al. | Structured volume decomposition via generalized sweeping | |
| CN112785526A (en) | Three-dimensional point cloud repairing method for graphic processing | |
| Hu et al. | Geometric feature enhanced line segment extraction from large-scale point clouds with hierarchical topological optimization | |
| Yeh et al. | 2.5 D cartoon hair modeling and manipulation | |
| Sharma et al. | Point cloud upsampling and normal estimation using deep learning for robust surface reconstruction | |
| Demir et al. | Proceduralization of buildings at city scale | |
| Li et al. | Advances in 3d generation: A survey | |
| Peng et al. | 3D hand mesh reconstruction from a monocular RGB image | |
| CN111191729B (en) | Three-dimensional object fusion feature representation method based on multi-modal feature fusion | |
| Liao et al. | Advances in 3D Generation: A Survey | |
| Baiocchi et al. | Artificial neural networks exploiting point cloud data for fragmented solid objects classification | |
| KR20200039050A (en) | Method for providing digital drawing and digital drawing providing device | |
| CN114842153A (en) | Method and device for reconstructing three-dimensional model from single two-dimensional wire frame diagram and electronic equipment | |
| Wattenberg et al. | A model of multi-scale perceptual organization in information graphics | |
| Collins et al. | Towards applicable Scan-to-BIM and Scan-to-Floorplan: an end-to-end experiment | |
| Yetiş | Auto-conversion from2D drawing to 3D model with deep learning | |
| Udayan | An analysis of reconstruction algorithms applied to 3d building modeling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |