摘要
充分利用形状蕴涵的语义信息进行三维形状的高层分析和理解是当前的热门话题。提出采用形状部件的上下文语义关系进行功能识别的方法,解决了当三维形状的几何特征和拓扑结构发生较大变化时形状部件的自动识别问题。首先,采用近似凸性分解技术将三维形状分割成具有独立语义的形状部件;然后,提出基于形状部件的上下文语义计算方法,并采用支持向量机实现形状部件自动识别。实验结果表明,相比于已有方法,可取得更高的部件匹配准确率和更低的分类错误率。
Making use of semantic information to achieve the high-level analysis and understanding is a hot issue currently. To address the problem of automatic recognition in the presence of significant geometric and topological variations,this paper proposed a 3D shape function recognition method by adopting the contextual relationship of shape parts. Firstly,it decomposed 3D shapes into the shape part sets with independent semantics and employed the technique of approximate convexity analysis.Then,it computed the contextual relationship of shape parts and on this basis,and adopted support vector machines to achieve the task of automatic recognition between shape parts. Experimental results show that the proposed method achieves higher the matching accuracy values and lower classification error rates,compared to the existing methods.
引文
[1] Fan Guoliang,Zhang Xin,Ding Meng. Gaussian process for human motion modeling:a comparative study[C]//Proc of IEEE International Workshop on Machine Learning for Signal Processing. Piscataway,NJ:IEEE Press,2011:1-6.
[2] Yu Tianshu,Wang Ruisheng. Scene parsing using graph matching on street-view data[J]. Computer Vision and Image Understanding,2016,145(4):70-80.
[3] Kim V G,Chaudhuri S,Guibas L,et al. Shape2Pose:human-centric shape analysis[J]. ACM Trans on Graphics,2014,33(4):article No. 120.
[4] Laga H,Mortara M,Spagnuolo M. Geometry and context for semantic correspondences and functionality recognition in man-made 3D shapes[J]. ACM Trans on Graphics,2013,32(5):article No. 150.
[5] Feng Jinlong,Liu Yushen,Gong Lianjie. Junction-aware shape descriptor for 3D articulated models using local shape-radius variation[J]. Signal Processing,2015,112(7):4-16.
[6] Léon V,Bonneel N,LavouéG,et al. Continuous semantic description of 3D meshes[J]. Computers&Graphics,2016,54(2):47-56.
[7] Kuang Zhenzhong,Li Zongmin,Lyu Qian,et al. Modal function transformation for isometric 3D shape representation[J]. Computers&Graphics,2015,46(2):209-220.
[8] Wang Jianyu,Yuille A. Semantic part segmentation using compositional model combining shape and appearance[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. Washington DC:IEEE Computer Society,2015:1788-1797.
[9]舒振宇,王鹏飞,于欣,等.基于局部稀疏表示的三维模型识别算法[J].计算机辅助设计与图形学学报,2014,26(11):1938-1947.(Shu Zhenyu,Wang Pengfei,Yu Xin,et al. 3D model recognition via local sparse representation[J]. Journal of Computer-Aided Design&Computer Graphics,2014,26(11):1938-1947.)
[10] Aubry M,Maturana D,Efros A A,et al. Seeing 3D chairs:exemplar part-based 2D-3D alignment using a large dataset of CAD models[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. Washington DC:IEEE Computer Society,2014:3762-3769.
[11]Daniela X,Berta C,Rubén F. A rule-based expert system for inferring functional annotation[J]. Applied Soft Computing,2015,35(10):373-385.
[12]Asafi S,Goren A,Cohen-Or D. Weak convex decomposition by linesof-sight[J]. Computer Graphics Forum,2013,32(5):23-31.
[13]Kaick O V,Fish N,Kleiman Y,et al. Shape segmentation by approximate convexity analysis[J]. ACM Trans on Graphics,2014,34(1):article No. 4.
[14]Li Pengfei,Dong Limin,Xiao Huachao,et al. A cloud image detection method based on SVM vector machine[J]. Neurocomputing,2015,169(12):34-42.
[15] Chen Xiaobai,Golovinskiy A,Funkhouser T. A benchmark for 3D mesh segmentation[J]. ACM Trans on Graphics,2009,28(3):article No. 73.
[16] COS. The shape coseg dataset[DB/OL].(2012). http://web. siat.ac. cn/yunhai/ssl/ssd. htm.