Supervised Learning of Salient 2D Views of 3D Models

We introduce a new framework for the automatic selection of the best views of 3D models based on the assumption that models belonging to the same class of shapes share the same salient features. The main issue is learning these features. We propose an algorithm for computing these features and their corresponding saliency value. At the learning stage, a large set of features are computed from every model and a boosting algorithm is applied to learn the classification function in the feature space. AdaBoost learns a classifier that relies on a small subset of the features with the mean of weak classifiers, and provides an efficient way for feature selection and combination. Moreover it assigns weights to the selected features which we interpret as a measure of the feature saliency within the class. Our experiments using the LightField (LFD) descriptors and the Princeton Shape Benchmark show the suitability of the approach to 3D shape classification and best-view selection for online visual browsing of 3D data collections.

[1]  Karthik Ramani,et al.  SVM-based Semantic Clustering and Retrieval of a 3D Model Database , 2005 .

[2]  Hans-Peter Kriegel,et al.  Nearest Neighbor Classification in 3D Protein Databases , 1999, ISMB.

[3]  Seungjin Choi,et al.  Supervised Learning , 2009, Encyclopedia of Biometrics.

[4]  Paul A. Viola,et al.  Boosting Image Retrieval , 2004, International Journal of Computer Vision.

[5]  Dejan V. VraniC.  An improvement of rotation invariant 3D-shape based on functions on concentric spheres , 2003, Proceedings 2003 International Conference on Image Processing (Cat. No.03CH37429).

[6]  Ali Shokoufandeh,et al.  Automated learning of model classifications , 2003, SM '03.

[7]  Karthik Ramani,et al.  A Probability-Based Unified 3D Shape Search , 2006, SAMT.

[8]  Jiebo Luo,et al.  Generalized Multiclass AdaBoost and Its Applications to Multimedia Classification , 2006, 2006 Conference on Computer Vision and Pattern Recognition Workshop (CVPRW'06).

[9]  Masayuki Nakajima,et al.  Spherical Wavelet Descriptors for Content-based 3D Model Retrieval , 2006, IEEE International Conference on Shape Modeling and Applications 2006 (SMI'06).

[10]  Thomas A. Funkhouser,et al.  Partial matching of 3D shapes with priority-driven search , 2006, SGP '06.

[11]  Silvia Biasotti,et al.  What’s in an image? , 2005, The Visual Computer.

[12]  David W. Jacobs,et al.  Mesh saliency , 2005, ACM Trans. Graph..

[13]  Thomas A. Funkhouser,et al.  Distinctive regions of 3D surfaces , 2007, TOGS.

[14]  Remco C. Veltkamp,et al.  SHREC2006: 3D Shape Retrieval Contest , 2006 .

[15]  M. Fatih Demirci,et al.  Selecting canonical views for view-based 3-D object recognition , 2004, Proceedings of the 17th International Conference on Pattern Recognition, 2004. ICPR 2004..

[16]  Ryutarou Ohbuchi,et al.  Unsupervised learning from a corpus for shape-based 3D model retrieval , 2006, MIR '06.

[17]  Hans-Peter Seidel,et al.  Towards Stable and Salient Multi-View Representation of 3D Shapes , 2006, IEEE International Conference on Shape Modeling and Applications 2006 (SMI'06).

[18]  Akihiro Yamamoto,et al.  Comparison of Dimension Reduction Methods for Database-Adaptive 3D Model Retrieval , 2007, Adaptive Multimedia Retrieval.

[19]  Thomas A. Funkhouser,et al.  Selecting Distinctive 3D Shape Descriptors for Similarity Retrieval , 2006, IEEE International Conference on Shape Modeling and Applications 2006 (SMI'06).

[20]  Robert E. Schapire,et al.  The Boosting Approach to Machine Learning An Overview , 2003 .

[21]  Nicu Sebe,et al.  Boosting contextual information in content-based image retrieval , 2004, MIR '04.

[22]  Ming Ouhyoung,et al.  On Visual Similarity Based 3D Model Retrieval , 2003, Comput. Graph. Forum.

[23]  Thomas A. Funkhouser,et al.  The Princeton Shape Benchmark , 2004, Proceedings Shape Modeling Applications, 2004..