Adaptive Feature Integration for Segmentation of 3D Data by Unsupervised Density Estimation

In this paper, a novel unsupervised approach for the segmentation of unorganized 3D points sets is proposed. The method derives by the mean shift clustering paradigm devoted to separate the modes of a multimodal density by using a kernel-based technique. Here, the attention is focused on the selection of the kernel bandwidth which typically strongly affects the level of accuracy of the segmentation results. In particular, a set of geometric features is computed from each 3D point of the given data. This set is projected onto a number of independent sub-spaces, each one associated to a different estimated feature, and overall forming a joint multidimensional (feature) space. In this space, we propose a method for selecting the best multidimensional kernel bandwidth in an automatic fashion, based on stability criteria. The final kernel considers each sub-space in an adaptive way in relation to the discrimination power of each feature, leading to accurate results when dealing with different types of 3D data

[1]  Z. Ésik,et al.  Research reports , 1977, Postgraduate medicine.

[2]  Keinosuke Fukunaga,et al.  Statistical Pattern Recognition , 1993, Handbook of Pattern Recognition and Computer Vision.

[3]  R. K. Hansen,et al.  A 3D Underwater Acoustic Camera — Properties and Applications , 1996 .

[4]  A WWW-Accessible Database for 3D Vision Research , 1998 .

[5]  Andrew R. Webb,et al.  Statistical Pattern Recognition , 1999 .

[6]  Dorin Comaniciu,et al.  The Variable Bandwidth Mean Shift and Data-Driven Scale Selection , 2001, ICCV.

[7]  Andrea Fusiello,et al.  Relaxing Symmetric Multiple Windows Stereo Using Markov Random Fields , 2001, EMMCVPR.

[8]  Andrea Fusiello,et al.  Reconstruction of complex environments by robust pre-aligned ICP , 2001, Proceedings Third International Conference on 3-D Digital Imaging and Modeling.

[9]  Andrea Fusiello,et al.  Disparity map restoration by integration of confidence in Markov random fields models , 2001, Proceedings 2001 International Conference on Image Processing (Cat. No.01CH37205).

[10]  U. Castellani,et al.  PRE-ALIGNED ICP FOR THE RECONSTRUCTION OF COMPLEX OBJECTS , 2001 .

[11]  Andrea Fusiello,et al.  Registration of Multiple Acoustic Range Views for Underwater Scene Reconstruction , 2002, Comput. Vis. Image Underst..

[12]  Dorin Comaniciu,et al.  Mean Shift: A Robust Approach Toward Feature Space Analysis , 2002, IEEE Trans. Pattern Anal. Mach. Intell..

[13]  Sylvain Petitjean,et al.  A survey of methods for recovering quadrics in triangle meshes , 2002, CSUR.

[14]  U. Castellani,et al.  VIEW SYNTHESIS FROM A SINGLE UNCALIBRATED IMAGE , 2002 .

[15]  Andrea Fusiello,et al.  Registration of very time-distant aerial images , 2002, Proceedings. International Conference on Image Processing.

[16]  Andrea Fusiello,et al.  Model Acquisition by Registration of Multiple Acoustic Range Views , 2002, ECCV.

[17]  Robert B. Fisher,et al.  Improving environment modelling by edge occlusion surface completion , 2002, Proceedings. First International Symposium on 3D Data Processing Visualization and Transmission.

[18]  Dorin Comaniciu,et al.  An Algorithm for Data-Driven Bandwidth Selection , 2003, IEEE Trans. Pattern Anal. Mach. Intell..

[19]  Manuele Bicego,et al.  Using hidden Markov models and wavelets for face recognition , 2003, 12th International Conference on Image Analysis and Processing, 2003.Proceedings..

[20]  Ilan Shimshoni,et al.  Mean shift based clustering in high dimensions: a texture classification example , 2003, Proceedings Ninth IEEE International Conference on Computer Vision.

[21]  Manuele Bicego,et al.  3D Face Recognition Using Stereoscopic Vision , 2003, Advanced Studies in Biometrics.

[22]  Linmi Tao,et al.  3D acoustic image segmentation by a RANSAC-based approach , 2003, Oceans 2003. Celebrating the Past ... Teaming Toward the Future (IEEE Cat. No.03CH37492).

[23]  Umberto Castellani,et al.  A versatile 3d modeling system based on skeleton extraction , 2003 .

[24]  Collections , 2003, Science.

[25]  Robert T. Collins,et al.  Mean-shift blob tracking through scale space , 2003, 2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2003. Proceedings..

[26]  Ariel Shamir.  Geodesic Mean Shift , 2004 .

[27]  U. Castellani,et al.  Efficient on-line mosaicing from 3D acoustical images , 2004, Oceans '04 MTS/IEEE Techno-Ocean '04 (IEEE Cat. No.04CH37600).

[28]  Linmi Tao,et al.  Robust 3D segmentation for underwater acoustic images , 2004, Proceedings. 2nd International Symposium on 3D Data Processing, Visualization and Transmission, 2004. 3DPVT 2004..

[29]  Andrea Fusiello,et al.  Fast model tracking with multiple cameras for augmented reality , 2004, VRST '04.

[30]  Vittorio Murino,et al.  UNDERWATER ENVIRONMENT MODELLING BY FAST 3D MOSAICING , 2005 .

[31]  M. Gobbi,et al.  Mosaicing of 3D sonar data sets - techniques and applications , 2005, Proceedings of OCEANS 2005 MTS/IEEE.

[32]  Hans-Peter Seidel,et al.  Feature sensitive mesh segmentation with mean shift , 2005, International Conference on Shape Modeling and Applications 2005 (SMI' 05).

[33]  Andrea Fusiello,et al.  Optimal Parameter Estimation for MRF Stereo Matching , 2005, ICIAP.

[34]  Umberto Castellani,et al.  Towards Information Visualization and Clustering Techniques for MRI Data Sets , 2005, AIME.

[35]  Umberto Castellani,et al.  Cancer area characterization by non-parametric clustering , 2006 .

[36]  Umberto Castellani,et al.  3D Data Segmentation Using a Non-Parametric Density Estimation Approach , 2006, Eurographics Italian Chapter Conference.

[37]  Umberto Castellani,et al.  Acoustic Range Image Segmentation by Effective Mean Shift , 2006, 2006 International Conference on Image Processing.

[38]  Umberto Castellani,et al.  Joint Reconstruction and Registration of a Deformable Planar Surface Observed by a 3D Sensor , 2007, Sixth International Conference on 3-D Digital Imaging and Modeling (3DIM 2007).

[39]  Andrea Fusiello,et al.  3D Objects Face Clustering using Unsupervised Mean Shift , 2007, Eurographics Italian Chapter Conference.

[40]  Alessandro Perina,et al.  Content visualization and management of geo-located image databases , 2008, CHI Extended Abstracts.

[41]  Adrien Bartoli,et al.  Coarse-to-fine low-rank structure-from-motion , 2008, 2008 IEEE Conference on Computer Vision and Pattern Recognition.

[42]  Alessandro Perina,et al.  Geo-located image analysis using latent representations , 2008, 2008 IEEE Conference on Computer Vision and Pattern Recognition.