<XML><RECORDS><RECORD><REFERENCE_TYPE>3</REFERENCE_TYPE><REFNUM>7433</REFNUM><AUTHORS><AUTHOR>Xiao,Y.</AUTHOR><AUTHOR>Siebert,J.P.</AUTHOR><AUTHOR>Werghi,N.</AUTHOR></AUTHORS><YEAR>2004</YEAR><TITLE>Topological Segmentation of Discrete Human Body Shapes in Different Postures Based on Geodesic Distance</TITLE><PLACE_PUBLISHED> The 17th International Conference on Pattern Recognition, Cambridge, UK </PLACE_PUBLISHED><PUBLISHER>IEEE Computer Society Press</PUBLISHER><LABEL>Xiao:2004:7433</LABEL><ABSTRACT> This paper presents a topological approach to segmenting clouds of discrete 3D points sampled on whole human body surfaces into subsets corresponding to primary body parts. In our approach, the Discrete Reeb Graph (DRG),an extension of the classical Reeb Graph to discrete datasets, is applied to abstract human body topology. Body part segmentation can be achieved by decomposing a constructed DRG according to its key critical nodes which are identified based on a criterion using topological and typographical information. In order to handle posture changes in human body scanning, a bending invariant function, namely Geodesic Distance, is employed as the Morse function to construct the DRG. The application of Geodesic distance also brings the independence of coordinate frame selection. We present a number of experiments conducted on both real body 3D scan samples and simulated datasets to demonstrate the validity of the approach </ABSTRACT><NOTES>accepted for publication</NOTES></RECORD></RECORDS></XML>