<XML><RECORDS><RECORD><REFERENCE_TYPE>3</REFERENCE_TYPE><REFNUM>7491</REFNUM><AUTHORS><AUTHOR>Mao,Z.</AUTHOR><AUTHOR>Siebert,J.P.</AUTHOR><AUTHOR>Cockshott,W.P.</AUTHOR><AUTHOR>Ayoub,A.F.</AUTHOR></AUTHORS><YEAR>2004</YEAR><TITLE>Constructing Dense Correspondences to Analyze 3D Facial Change</TITLE><PLACE_PUBLISHED>The 17th International Conference on Pattern Recognition, Cambridge, UK </PLACE_PUBLISHED><PUBLISHER>Dept of Computing Science, University of Glasgow</PUBLISHER><LABEL>Mao:2004:7491</LABEL><ABSTRACT> Although corresponding landmarks and curves have been used widely to analyze shape or shape change, neither of these is able to provide a full 3D description of a surface. Constructing dense correspondences between compared 3D surfaces is a natural extension of the landmark-based approach. It presents us with much richer 3D information and hence enables us to properly evaluate 3D objects and even construct a 3D PDM ([4]). In this paper, we will present a robust method to construct dense correspondences between different 3D models. Compared to the approach in [7], our method exhibits greater accuracy and consistency </ABSTRACT><NOTES>accepted for application</NOTES></RECORD></RECORDS></XML>