Publications by year (partial list only)
- My Google Scholar (whole list)
- Publications by journal and conference
- Technical reports
- Invited talks, tutorials, and courses
Journal and Refereed Conference Proceedings
2016
Risheng Liu*, Guangyu Zhong, Junjie Cao, Zhouchen Lin, Shiguang Shan, Zhongxuan Luo. Learning to Diffuse: A New Perspective to Design PDEs for Visual Analysis. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). Accepted, 2016. [Project page |] (SCI, IF: 5.781) (CFC A)
Partial differential equations (PDEs) have been used to formulate image processing for several decades. Generally, a PDE system consists of two components: the governing equation and the boundary condition. In most previous works, both of them are generally designed by people using mathematical skills. So these PDEs highly depend on good insight and deep domain knowledge to particular vision tasks. Moreover, the predefined, fixed-form PDEs may not be able to describe the complex structure of the visual data, especially the labeling information and the discriminative distribution. To address above issues, we propose a new PDE framework, named learning-to-diffuse (LTD), to adaptively design the governing equation and the boundary condition of a diffusion PDE system for various vision tasks on different types of visual data. To our best knowledge, the problems considered in this paper (i.e., saliency detection and object tracking) have never been addressed by PDE models before. Experimental results on various challenging benchmark databases show the superiority of LTD against existing state-of-the-art methods for all the tested visual analysis tasks
Junjie Cao, Jie Zhang, Zhijie Wen*, Nannan Wang, Xiuping Liu*, Fabric Defect Inspection using Prior Knowledge Guided Least Squares Regression. Multimedia Tools and Applications. Accepted, 2015. (SCI, IF: 1.346) (CFC C)
This paper proposes an unsupervised model to inspect various detects in fabric images with diverse textures. A fabric image with defects is usually composed of a relatively consistent background texture and some sparse defects, which can be represented as a low-rank matrix plus a sparse matrix in a certain feature space. The process is formulated as a least squares regression based subspace segmentation model, which can be solved efficiently. Instead of considering only the feature space' s global structure, a local prior is also learnt from the image itself and incorporated with it seamlessly by the proposed subspace segmentation model to guide and improve the segmentation.
Xiuping Liu, Pingping Tao, Junjie Cao*, He Chen, Changqing Zou. Mesh saliency detection via double absorbing Markov chain in feature space. The Visual Computer. Accepted, 2015. (SCI, IF: 0.957) (CFC C)
We propose mesh saliency detection approach using an absorbing Markov chain. Unlike most of the existing methods based on some center-surround operator, our method employs feature variance to obtain insignicant regions and considers both background and foreground cues. Firstly, we partition an input mesh into a set of segments and then each segment is further over segmented into patches based on Zernike coefficients. Afterwards, some background patches are selected by computing feature variance within the segments. Secondly, the absorbed time of each node is calculated via absorbing Markov chain with the background patches as absorbing nodes, which gives a preliminary saliency measure. Thirdly, a refined saliency result is generated in a similar way but with foreground nodes extracted from the preliminary saliency map as absorbing nodes, which inhibits the background and efficiently enhances salient foreground regions. Finally, a Laplacian based smoothing procedure is utilized to spread the patch saliency to each vertex. Experimental results demonstrate that our scheme performs favourably against the state-of-the-art approaches.
2015
Ibraheem Alhashim, Kai Xu, Yixin Zhuang, Junjie Cao, Patricio Simari, and Hao Zhang, "Deformation-Driven Topology-Varying 3D Shape Correspondence", ACM Trans. on Graphics (Special Issue of SIGGRAPH Asia), Vol. 34, No. 6, Article 236, 2015. [PDF (20 MB) | PDF reduced (12 MB) | Project page | bibtex] (CFC A)
We present a deformation-driven approach to topology-varying 3D shape correspondence. In this paradigm, the best correspondence between two shapes is the one that results in a minimal-energy, possibly topology-varying, deformation that transforms one shape to conform to the other while respecting the correspondence. Our deformation model allows both geometric and topological operations such as part split, duplication, and merging ...
Xiuping Liu, Jie Zhang, Junjie Cao*, Bo Li, Ligang Liu. Quality Point Cloud Normal Estimation by Guided Least Squares Representation. Computers & Graphics (Special Issue of SMI 2015), 2015, 51, 106-116. (SCI, IF: 1.03) (CFC C).
1) A guided subspace clustering framework and its theoretical analysis are presented.
2) SGLRS is proposed as a special case of the framework.
3) A rapid algorithm with a natural parallelism is devised for SGLRS.
4) A quality normal estimation method is proposed based on SGLRS.
5) A subspace structure propagation algorithm is designed to speed up the estimation.
2) SGLRS is proposed as a special case of the framework.
3) A rapid algorithm with a natural parallelism is devised for SGLRS.
4) A quality normal estimation method is proposed based on SGLRS.
5) A subspace structure propagation algorithm is designed to speed up the estimation.
Pingping Tao, Junjie Cao*, Shuhua Li, Xiuping Liu, Ligang Liu. Mesh Saliency via Ranking Unsalient Patches in a Descriptor Space. Computers & Graphics (Special Issue of SMI 2014), 2015, 46, 264-274. (SCI, IF: 1.029) (CFC C).
1) A simple mesh saliency detection method via manifold ranking is proposed.
2) We select some unsalient background patches as queries to achieve more robust results.
3) Ranking in the descriptor space of the patches reveals the saliency patches independent of their locations and cardinality.
4) Our method achieves faithful results using only single scale descriptor.
5) Our algorithm is comparable with state-of-the-art methods.
2) We select some unsalient background patches as queries to achieve more robust results.
3) Ranking in the descriptor space of the patches reveals the saliency patches independent of their locations and cardinality.
4) Our method achieves faithful results using only single scale descriptor.
5) Our algorithm is comparable with state-of-the-art methods.
Xiuping Liu, Jie Zhang, Risheng Liu, Bo Li, Jun Wang, Junjie Cao*. Low-rank 3D Mesh Segmentation and Labeling with Structure Guiding. Computers & Graphics (Special Issue of SMI 2014), 2015, 46, 99-109. (SCI, IF: 1.029) (CFC C).
1) A low-rank representation model with structure guiding is designed to label 3D mesh.
2) The tedious pre-training step existing in the data-driven approaches is eliminated.
3) Just using a few examples, a test mesh can be labeled faithfully.
4) Our method is robust to the seriously mislabeled examples.
5) Correct labeling is got even if given examples include multiple object categories.
2) The tedious pre-training step existing in the data-driven approaches is eliminated.
3) Just using a few examples, a test mesh can be labeled faithfully.
4) Our method is robust to the seriously mislabeled examples.
5) Correct labeling is got even if given examples include multiple object categories.
Xiuping Liu, Shuhua Li, Risheng Liu, Jun Wang, Hui Wang*, Junjie Cao*. Properly-constrained Orthonormal Functional Maps for Intrinsic Symmetries. Computers & Graphics (Special Issue of SMI 2014), 2015, 46, 198-208. (SCI, IF: 1.029) (CFC C).
1) A robust global intrinsic symmetry method based on functional maps is proposed.
2) We employ an optimization with orthogonality constraints and solve it directly.
3) An reliable symmetry-invariant set is extracted to establish initial constraints.
4) A voting scheme is proposed to establish more regional constraints.
5) A considerable amount of comparisons with the state-of-the-art are provided.
2) We employ an optimization with orthogonality constraints and solve it directly.
3) An reliable symmetry-invariant set is extracted to establish initial constraints.
4) A voting scheme is proposed to establish more regional constraints.
5) A considerable amount of comparisons with the state-of-the-art are provided.
Guangyu Zhong, Risheng Liu, Junjie Cao*, Zhixun Su. A generalized nonlocal mean framework with object-level cues for saliency detection. The Visual Computer. 2015, ?, 1-13. (SCI, IF: 0.957) (CFC C) [Project page]
Nonlocal mean (NM) is an efficient method for many low-level image processing tasks. However, it is challenging to directly utilize NM for saliency detection. This is because that conventional NM method can only extract the structure of the image itself and is based on regular pixel-level graph. However, saliency detection usually requires human perceptions and more complex connectivity of image elements. In this paper, we propose a novel generalized nonlocal mean (GNM) framework with the object-level cue which fuses the low-level and high-level cues to generate saliency maps.
Risheng Liu*, Guangyu Zhong, Junjie Cao, Zhixun Su. Diffuse Visual Attention for Saliency Detection. Journal of Electronic Imaging. 2015, 24(1), 13-23. (IF=0.85, SCI)
Zhiyang Li, Wenyu Qu*, Junjie Cao, Heng Qi, Milos Stojmenovic. ECDS: An effective shape signature using electrical charge distribution on the shape, Pattern Recognition, 2015, 48(2), 402-410. (SCI, IF:2.584) (CFC B)
1) A shape signature ECDS is proposed based on the electrical charge distribution.
2) ECDS measures the local amount of charge but computed in a global manner.
3) ECDS is articulation insensitive and effective at capturing partial structures.
4) The sum of ECDS remains constant during the process of electrical equilibrium.
5) ECDS compares well with other descriptors in various shape recognition tasks.
2) ECDS measures the local amount of charge but computed in a global manner.
3) ECDS is articulation insensitive and effective at capturing partial structures.
4) The sum of ECDS remains constant during the process of electrical equilibrium.
5) ECDS compares well with other descriptors in various shape recognition tasks.
Junjie Cao, Shuhua Li, Xiuping Liu*. Functional Maps with Well Spread-out Constraints. JOURNAL OF INFORMATION AND COMPUTATIONAL SCIENCE . 2015. 12 (12) : 4531- 4538. (EI)
Finding an isometric map between two intrinsically symmetric surfaces is a fundamental problem in computer graphics. However, the deviations from perfect isometry make this problem challenging to the state-of-the-art
methods. In this paper, we propose an automatic and robust algorithm based on functional maps for finding nearly isometric maps. The challenge is to extract sufficient and reliable constraints for functional maps. Based on the observation that symmetry axis curves and shape extremities are stable and well-spaced features, we extract and match symmetry axis curves and feature points over shape extremities. These sparse point correspondences are then extrapolated to obtain well spread-out regional constraints. Experimental results on SCAPE data set show that our method is superior to the state-of-the-art methods.
methods. In this paper, we propose an automatic and robust algorithm based on functional maps for finding nearly isometric maps. The challenge is to extract sufficient and reliable constraints for functional maps. Based on the observation that symmetry axis curves and shape extremities are stable and well-spaced features, we extract and match symmetry axis curves and feature points over shape extremities. These sparse point correspondences are then extrapolated to obtain well spread-out regional constraints. Experimental results on SCAPE data set show that our method is superior to the state-of-the-art methods.
Baochang Han, Junjie Cao, Zhixun Su*. Automatic Point Clouds Registration Based on Regions. Journal of Computer-Aided Design & Computer Graphics, 27(2), 313-319, 2015. (EI)
We present a region-based algorithm for the automatic registration of 3D point clouds. Most existed algorithms align the two scans globally, making them unsuitable when the overlapping ratio is low or the input shapes do not have strong features. We notice that rigid transform is low-dimensional and two overlapped regions of the point clouds are enough to recover it. Thus, we align each pair of regions directly, and then solve an energy optimization to obtain the global transform from a series of region registrations by introducing confidence term and consistency term. Finally, sparse ICP algorithm is used for refinement. Experiments show that under premise of robustness to noise and outliers, our algorithm can align scans with less overlapping ratio.
Hui WANG, Junjie CAO, Xiuping LIU*, Yue WANG. Guided Random Walk Segmentation for Images with Textures. Journal of Computational Information Systems. 2015, 8(1), 1-13. (EI).
Random walk image segmentation is one of the state-of-art supervised methods. However, it fails for images containing high contrast and fine-scale textures. In this paper, we extend the original random walk for such images. The key of our extension is that a guided image based on patch shift is introduced to control the segmentation process. The additional guided image computation has a smaller computation cost compared with the original random walk. Furthermore, the proposed method inherits advantages of the original random walk of image segmentation. Experimental results demonstrate the proposed guided random walk segmentation achieves high quality results.
Hui Wang, Junjie Cao, Xiuping Liu*, Jianmin Wang, Tongrang Fan, Jianping Hu. Least-square images for edge-preserving smoothing, Computational Visual Media, 2015, 1(1): 27-35.
In this paper, we propose least-squares images (LS-images) as a basis for a novel edge-preserving image smoothing method. The LS-image requires the value of each pixel to be a convex linear combination of its neighbors, i.e., to have zero Laplacian, and to approximate the original image in a least-squares sense. The edge-preserving property inherits from the edge-aware weights for constructing the linear combination. Experimental results demonstrate that the proposed method achieves high quality results compared to previous state-of-the-art works. We also show diverse applications of LS-images, such as detail manipulation, edge enhancement, and clip-art JPEG artifact removal.
2014
Risheng Liu*, Junjie Cao,
Zhouchen Lin, Shiguang Shan. Adaptive Partial Differential Equation Learning
for Visual Saliency Detection. IEEE International Conference on Computer Vision
and Pattern Recognition (CVPR), 2014: 3866-3873. (Oral, Accept rate: 5.75%) (CFC A) [PDF]
Instead of designing PDEs with fixed formulation and boundary condition, this paper proposes a novel framework for adaptively learning a PDE system from an image for visual saliency detection. We assume that the saliency of image elements can be carried out from the relevances to the saliency seeds (i.e., the most representative salient elements). In this view, a general Linear Elliptic System with Dirichlet boundary (LESD) is introduced to model the diffusion from seeds to other relevant points.
Ibraheem Alhashim, Honghua Li, Kai Xu, Junjie Cao, Rui Ma, and Hao Zhang, "Topology-Varying 3D Shape Creation via Structural Blending", ACM Trans. on Graphics (Special Issue of SIGGRAPH), 33(4), Article 158, 2014. (SCI, IF: 3.361) (CFC A) [Project page | PDF | bibtex]
We introduce an algorithm for generating novel 3D models via topology-varying shape blending. Given a source and a target shape, our method blends them topologically and geometrically, producing continuous series of in-betweens as new shape creations. The blending operations are defined on a spatio-structural graph composed of medial curves and sheets. Such a shape abstraction is structure-oriented, part-aware, and facilitates topology manipulations. Fundamental topological operations including split and merge are realized by allowing one-to-many correspondences between the source and the target ...
Jian Liu, Junjie Cao*, Xiuping Liu, Jun Wang, XiaoChao Wang, Xiquan Shi. Mendable consistent orientation of point clouds. Computer-Aided Design, 2014, 55: 26-36. (SCI, IF:1.264) (CFC B) [PDF|Code|bibtex]
• An orientation-benefit normal estimation method is proposed.
• We use multi-sources normal propagation to achieve more consistent orientation.
• Propagation sources are extracted automatically to alleviate the manual work.
• A considerable amount of comparisons with state-of-the-art show that our method improves existing methods.
• We use multi-sources normal propagation to achieve more consistent orientation.
• Propagation sources are extracted automatically to alleviate the manual work.
• A considerable amount of comparisons with state-of-the-art show that our method improves existing methods.
Shengfa Wang, Yu Cai, Zhiling Yu, Junjie Cao, Zhixun Su. Normal-controlled Coordinates Based Feature-preserving Mesh Editing. Multimedia Tools and applications, 2014, 71(2): 607-622. (SCI, IF:1.014) (CFC C)
As a local shape descriptor, normal-controlled coordinates (NCC) are
well defined on the boundary vertices of open meshes, and proven being always parallel with the corresponding vertex normals, which means no tangential drift appears in various editing operations. ...
well defined on the boundary vertices of open meshes, and proven being always parallel with the corresponding vertex normals, which means no tangential drift appears in various editing operations. ...
Zhijie Wen*, Junjie Cao, Xiuping Liu, Shihui Ying. Fabric defects detection using adaptive wavelets. International Journal of Clothing Science and Technology. 2014, 26(3): 202 - 211. (SCI, IF: 0.458)
Li, Zhiyang; *Qu, Wenyu; Xu, Yujie; Cao, Junjie; Su, Zhixun. Efficient shape representation and retrieval in large database. International Journal of Computer Systems Science & Engineering, 29(4), pp 275-284, 2014/7. (SCI, IF: 0.235)
Yuxia Song, Junjie Cao, Risheng Liu, Guangyu Zhong, Zhixun Su*. Saliency detection via Ranking with Reconstruction Error, Journal of Information & Computational Science 11:13 (2014) 4467–4476. (EI)
Shengfa Wang, Junjie Cao, Hui Wang*, Baochang Han, Bo Li, Zhixun Su. Primary Correspondences between Intrinsically Symmetrical Shapes. Journal of Information and Computational Science, 2014, 11(9): 2975-2982. (EI)
Nannan Li, Junjie Cao, Peng Wang, Hui Wang, Zhixun Su*, Bo Li. Reweighted Global Bilateral Filtering Based on Normal Regularization. Journal of Computer Aided Design & Computer Graphics, 2014, 26 (3): 370-377. (EI)
2013
Jie Zhang, Junjie Cao*, Xiuping Liu, Jun Wang, Jian Liu, Xiquan Shi, "Point cloud normal estimation via low-rank subspace clustering", Computer & Graphics (SMI 2013), 2013, 37(6): 697-706. (SCI, IF: 1.029) (CFC C) [PDF|Code|bibtex]
We present a robust normal estimation algorithm based on the low-rank subspace clustering technique. The main idea is based on the observation that compared with the points around sharp features, it is relatively easier to obtain accurate normals for the points within smooth regions. ...
Honghua Li, Hao Zhang, Yanzhen Wang, Junjie Cao, Ariel Shamir, Daniel Cohen-Or, "Curve Style Analysis in a Set of Shapes," Computer Graphics Forum, 2013, 32(6), 77-88. (SCI, IF:1.638) (CFC B) [PDF | bibtex]
We pose the open question "how to extract styles from geometric shapes?" and address one instance of the problem. Specifically, we present an unsupervised algorithm for identifying curve styles in a set of shapes ...
Wang, Jun; Yu, Zeyun; Zhu, Weidong; Cao, Junjie, "Feature-Preserving Surface Reconstruction from Unoriented, Noisy Point Data", Computer Graphics Forum, 2013, 32(1), 164-176. (SCI, IF:1.638) (CFC B) [PDF]
We propose a robust method for surface mesh reconstruction from 3D unorganized, unoriented, noisy and outlier-ridden point data. ...
Jun Wang, Kai Xu, Ligang Liu, Junjie Cao, Shengjun Liu, Zeyun Yu, Xianfeng Gu. Consolidation of Low-quality Point Clouds from Outdoor Scenes. Computer Graphics Forum (SGP 2013), 32(5), 207-216. (SCI, IF:1.638) (CFC B) [PDF]
This paper presents a robust consolidation algorithm for low-quality point data from outdoor scenes, which essentially consists of two steps: 1) outliers filtering and 2) noise smoothing. ...
Zhenzhen Zhang, Junjie Cao*, Guangyu Zhong, Wangyi Liu, Zhixun Su. Object Level Image Saliency by Hierarchical Segmentation. ICIP 2013. (CFC C)
We propose a saliency detection method with the aim to highlight objects as a whole and distinguish objects with different saliency levels. It combines the bottom-up approach and top-down approach via two nested levels of hierarchical segmentations - the coarse level objects and fine level details. ...
Riming Sun, Junjie Cao, Fangfang Qiu, Ruru Hao, Zhixun Su*. Geometry Image with Higher Compressibility. Journal of Information and Computational Science, 2013, 10 (16): 5135-5143. (EI)
Riming Sun, Shengfa Wang*, Junjie Cao, Bo Li, Zhixun Su. An Adapted Parameterization for Smooth Geometry Images. The 13th International Conference on Computer-Aided Design and Computer Graphics (CAD/Graphics), 2013, 156-163. (EI).
2012
Hui Wang, Zhixun Su, Junjie Cao, Ye Wang, and Hao Zhang, "Empirical Mode Decomposition on Surfaces," Graphical Models (Special Issue of GMP), 2012, 74(4): 173–183. (SCI) [PDF |bibtex]
Xiaochao Wang, Xiuping Liu, Linfa Lu, Baojun Li, Junjie Cao, Baocai Yin, Xiquan Shi, Automatic hole-filling of CAD models with feature-preserving, Computers & Graphics, 2012, 36(2): 101-110. (SCI) [PDF|bibtex]
In this paper, we propose an automatic hole-filling method, particularly for recovering missing feature curves and corners. We first extract the feature vertices around a hole of a CAD model and classify them into different feature sets. These feature sets are then automatically paired, using ordered double normals, Gaussian mapping and convex/concave analysis, to produce missing feature curves. Additionally, by minimizing a newly defined energy, the missing corners can be efficiently recovered as well. The hole is consequently divided into simple sub-holes according to the produced feature curves and recovered corners. Finally, each sub-hole is filled by a modified Advancing Front method individually. The experiments show that our approach is simple, efficient, and suitable for CAD systems.
Xiaochao Wang, Junjie Cao, Xiuping Liu, Baojun Li, Xiquan Shi, Yizhen Sun. Feature detection on triangular meshes via neighbor supporting. Journal of Zhejiang University-SCIENCE C (Computers & Electronics), 2012, 13(6), 440-451. (SCI) [PDF|bibtex|code]
In this paper, we propose a robust method for detecting features on triangular meshes by combining normal tensor voting with Neighbor Supporting. Our method contains two stages: feature detection and feature refinement. First, the normal tensor voting method is modified to detect the initial features, which may include some pseudo features. Then, at the feature refinement stage, a novel salient measure deriving from the idea of Neighbor Supporting is developed. Benefitting from the integrated reliable salient measure feature, pseudo features can be effectively discriminated from the initially detected features and removed. Compared to previous methods based on differential geometry property, the main advantage of our method is that it not only detects sharp features, but weak features as well. Numerical experiments show that our algorithm is robust, effective, and can produce more accurate results.
2011
Junjie Cao*, Ying He, Zhiyang Li, Xiuping Liu, Zhixun Su. Orienting Raw Point Sets by Global Contraction and Visibility Voting. Computer & Graphics (short paper of SMI 2011). (SCI) [PDF|PPT|Poster|bibtex]
We
present a global orienter for defective raw point clouds. It seamlessly
combines constrained Laplacian smoothing and visibility voting. It is
simple and easy to implement. The noisy point sets with thin sharp
features and holes can be robustly oriented. Implicit reconstruction
schemes recover high quality surfaces with the help of it.
Hui Wang, Hongyin Chen, zhixun Su, Junjie Cao, Fengshan Liu, Xiquan Shi. Versatile surface detail editing via Laplacian coordinates. The Visual Computer, 2011. (SCI) [PDF|bibtex]
We present a versatile detail editing approach for triangular meshes based on filtering the Laplacian
coordinates. The proposed detail editing method includes not only feature preserving smoothing but also enhancing. Furthermore, the proposed approach allows interactive editing of some user-specified frequencies and regions. Experimental results demonstrate that our method is much more versatile and faster than the existing methods.
coordinates. The proposed detail editing method includes not only feature preserving smoothing but also enhancing. Furthermore, the proposed approach allows interactive editing of some user-specified frequencies and regions. Experimental results demonstrate that our method is much more versatile and faster than the existing methods.
Yuandi Zhao, Junjie Cao, Zhixun Su, Zhiyang Li. Efficient Reconstruction of Non-simple Curves. Journal of Zhejiang University - Science C, 2011. (SCI)( [PDF|bibtex]
We reconstruct curves with self-intersections and multiple parts from unorganized strip-shaped points, which may have different local shape scales and sampling densities.
We first extract an initial curve, a graph composed of polylines, to model the different structures of the points. Then a least-squares optimization is used to improve the geometric approximation. Our algorithm produces faithful results for points sampled from non-simple curves without pre-segmenting them. Experiments on many simulated and real data demonstrate the efficiency of our method, and more faithful curves are reconstructed compared to other existing methods.
We first extract an initial curve, a graph composed of polylines, to model the different structures of the points. Then a least-squares optimization is used to improve the geometric approximation. Our algorithm produces faithful results for points sampled from non-simple curves without pre-segmenting them. Experiments on many simulated and real data demonstrate the efficiency of our method, and more faithful curves are reconstructed compared to other existing methods.
Xiaochao Wang, Junjie Cao, Baojun Li, Xiuping Liu. AdvancingFront Method in Triangular Meshes Hole-filling Application. Journal of Computer-Aided Design & Computer Graphics, 2011,23(6):1048-1054. (EI) (Chinese version, Main algorithm in English)
In order to restore the missing shape of the hole in triangular meshes, especially big ones locating at high curved region, a hole-filling algorithm based on advancing front method is proposed in this paper. After detecting the boundary of the hole, normals of the boundary vertices are well estimated. Combining with the Laplacian coordinate, the boundary vertices are classified into two types: concave and convex. Then, based on the normal, the concavity-convexity feature of each boundary vertex and a proper adjustment parameter, optimal vertices are carefully computed and new triangles are created to fill holes. Many experimental results show that our method has powerful ability to recover the missing shape with high quality triangular mesh for even big holes located at the high curved regions. Without post-processing, such as refinement and smoothing, the hole-filing meshes obtained by our method interpolate the shape and have consistent mesh distribution and smooth transition with the surrounding meshes.
2010
Junjie Cao*, Andrea Tagliasacchi, Matt Olson, Hao Zhang, and Zhixun Su, "Point Cloud Skeletons via Laplacian-Based Contraction," Proc. of IEEE Shape Modeling International, 187-19, 2010. [PDF | PPT | Poster| Code | Project page | bibtex]
We develop a contraction operation that is designed to work on generalized discrete geometry data, particularly point clouds, via local Delaunay triangulation and topological thinning. Our approach is robust to noise and can handle moderate amounts of missing data, allowing skeleton-based manipulation of point clouds without explicit surface reconstruction. By avoiding explicit reconstruction, we are able to perform skeleton-driven topology repair of acquired point clouds in the presence of large amounts of missing data. In such cases, automatic surface reconstruction schemes tend to produce incorrect surface topology. We show that the curve skeletons we extract provide an intuitive and easy-to-manipulate structure for effective topology modification, leading to more faithful surface reconstruction.
Junjie Cao*, Zhixun Su, Xiuping Liu, Haichuan Bi. Measured Boundary Parameterization Based on Poisson Equation. Journal of Zhejiang University - SCIENCE C, 11(3), 187-198, 2010. [PDF | Project page | bibtex] (SCI)
One major goal of mesh parameterization is to minimize the conformal distortion. Measured boundary parameterizations focus on lowering the distortion by setting the boundary free with the help of distance from a center vertex to all the boundary vertices. Hence these parameterizations strongly depend on the determination of the center vertex. In this paper, we introduce two methods to determine the center vertex automatically. Both of them can be used as necessary supplements to the existing measured boundary methods to minimize the common artifacts as a result of the obscure choice of the center vertex. In addition, we propose a simple and fast measured boundary parameterization method based on the Poisson’s equation. Our new approach generates less conformal distortion than the fixed boundary methods. It also generates more regular domain boundaries than other measured boundary methods. Moreover, it offers a good tradeoff between computation costs and conformal distortion compared with the fast and robust angle based flattening (ABF++).
2009
Zhixun Su, Yuandi Zhao, Junjie Cao. "An adaptive method for constrained texture mapping". Journal of Computer-Aided Design & Computer Graphics, 21(12), 1722-1728, 2009. (EI: 20100312640445)
Most of the traditional constrained texture mapping methods meet the specified constrained conditions at the cost of sacrificing the parameterizationps conformality. In this paper, a new adaptive constrained texture mapping method is proposed. The constrained parameterizations are computed through iteratively minimizing a weighted energy covering both conformal term and constraint term. In order to make them in balance, the weights in the energy are updated step by step. In one iteration step, the weight of each constrained vertex is dependent on a monotone function of the distance from current position of the vertex to its specified location. By this method, the conformality can be well preserved while the constrained conditions are satisfied. Experimental results show that a good texture mapping can be taken quickly and effectively.
Zhixun Su, Hui Wang, Junjie Cao. "Mesh denoising based on differential coordinates". IEEE International Conference on Shape Modeling and Applications (SMI2009),1-6, 2009. (SCI Conference: 000272068600001, EI PageOne, ISTP) [PDF]
We propose a novel triangle mesh denoising method based on the differential coordinates. The proposed approach consists of the application of the mean filter to differential coordinates of the mesh and the reconstruction of mesh vertices’ Cartesian coordinates to make them fit to the modified differential coordinates. The presented method is simple, stable and able to effectively remove large noise. Experimental results demonstrate that the proposed Mesh Mean Filter does not cause surface shrinkage and shape distortion during the denoising process, and preserves geometric detail features to a certain extent.