PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Scale Invariant Features for 3D Mesh Models
Yosi Keller and Tal Darom
IEEE Transactions on Image Processing 2010.


In this paper we present a framework for detecting interest points in three-dimensional meshes and computing their corresponding descriptors. For that we propose an intrinsic scale detection scheme per interest point, and utilize it to derive two scale invariant local features for mesh models. First, we present the Scale Invariant Spin Image local descriptor, that is a scale-invariant formulation of the Spin Image descriptor [11]. Secondly, we adapt the SIFT feature [16] to mesh data by representing the vicinity of each interest point as a depth map and estimating its dominant angle using PCA to achieve rotation invariance. The proposed features are experimentally shown to be robust to scale changes and partial mesh matching, and compared favorably to other local mesh features on the SHREC’11 testbed. We also applied the proposed local features to mesh retrieval using the Bag-of-Features approach and achieved state-of-the-art retrieval accuracy.

EPrint Type:Article
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Machine Vision
ID Code:7679
Deposited By:Yosi Keller
Deposited On:17 March 2011