摘要
数据驱动的有监督联合分割可以通过先验知识的学习,达到更精确的分割与标注要求。然而,目前的有监督分割方法大多需要耗费大量的训练时间,不利于大规模数据集的扩展。为了提高学习效率,提出一种基于极限学习机同时对面片和网格边进行训练的快速的三维形状分割和标注方法。进而通过图割优化进行分割边缘的平滑和优化,得到最终的标注结果。实验结果表明,在三维形状的分割和标注过程中,该方法学习快速,且可以达到较高的分割精度和视觉效果。
Data-driven supervised co-segmentation can achieve more accurate segmentation and labeling requirements based on prior knowledge. However, most of supervised methods are extremely time-consuming and difficult to scale up to large data set. The fast 3D shape segmentation and labeling learning method via extreme learning machine is provided,which trains facets and edges simultaneously. Based on that, graph-cut is adopted to smooth and optimize the segmentation boundaries. The experimental results show that this method can learn quickly and achieve high segmentation accuracy and visual effect in the process of 3D shape segmentation and labeling.
引文
[1]胡事民,杨永亮,来煜坤.数字几何处理研究进展[J].计算机学报,2009,32(8):1451-1469.
[2]Li H Y,Sun Z X,He M,et al.A mesh watermarking method based on local roughness analysis[C]//Proceedings of IEEE International Conference on Software Engineering and Service Science,2015:379-383.
[3]Guo X K,Lin J C,Xu K,et al.Custom Cut:on-demand extraction of customized 3D parts with 2D sketches[J].Computer Graphics Forum,2016,35(5):89-100.
[4]Guo X K,Lin J C,Xu K,et al.Creature grammar for creative modeling of 3D monsters[J].Graphical Models,2014,76(5):376-389.
[5]Xu K,Kim V G,Huang Q,et al.Data-driven shape analysis and processing[J].Computer Graphics Forum,2016:1-32.
[6]Guo K,Zou D,Chen X.3D mesh labeling via deep convolutional neural networks[J].ACM Transactions on Graphics,2015,35(1):1-12.
[7]Kalogerakis E,Hertzmann A,Singh K.Learning 3d mesh segmentation and labeling[J].ACM Transactions on Graphics,2010,29:1-12.
[8]Xie Z,Xu K,Liu L,et al.3D shape segmentation and labeling via extreme learning machine[J].Computer Graphics Forum,2014,33(5):85-95.
[9]van Kaick O,Tagliasacchi A,Sidi O,et al.Prior knowledge for part correspondence[J].Computer Graphics Forum,2011,30(2):553-562.
[10]Hu R,Fan L,Liu L.Co-segmentation of 3d shapes via subspace clustering[J].Computer Graphics Forum,2012,31(5):1703-1713.
[11]Sidi O,van Kaick O,Kleiman Y,et al.Unsupervised cosegmentation of a set of shapes via descriptor-space spectral clustering[J].ACM Transactions on Graphics,2011,30(6):1-10.
[12]Lv J,Chen X,Huang J,et al.Semi-supervised mesh segmentation and labeling[J].Computer Graphics Forum,2012,31(7):2241-2248.
[13]Wang Y,Asafi S,van Kaick O,et al.Active co-analysis of a set of shapes[J].ACM Transactions on Graphics,2012,31(6):1-10.
[14]Benhabiles H,LavouéG,Vandeborre J P,et al.Learning boundary edges for 3d-mesh segmentation[J].Computer Graphics Forum,2011,30(8):2170-2182.
[15]Boykov Y,Veksler O,Zabih R.Fast approximate energy minimization via graph cuts[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2001,23(11):1222-1239.
[16]Huang G B,Zhou H,Ding X,et al.Extreme learning machine for regression and multiclass classification[J].IEEE Transactions on Systems,Man,and Cybernetics,2012,42(2):513-529.
[17]Shapira L,Shalom S,Shamir A,et al.Contextual part analogies in 3D objects[J].International Journal of Computer Vision,2010,89:309-326.
[18]Hilaga M,Shinagawa Y,Kohmura T,et al.Topology matching for fully automatic similarity estimation of 3d shapes[C]//Proceedings of the 28th Annual Conference on Computer Graphics and Interactive Techniques,2001:203-212.
[19]Belongie S,Malik J,Puzicha J.Shape matching and object recognition using shape contexts[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2002,24(4):509-522.