Due to the recent improvements in laser scanning technology, the acquisition of 3D models by 3D digitizing, now is commonplace. Applications of emerging relevance are augmented reality using digitized 3D models, 3D shape retrieval, and the creation of digital archives for all kind of purposes, e.g. recording cultural heritage and reverse engineering. The World Wide Web enables access to these digital archives and desktop computers now have the power to process and display huge 3D data sets. Inspired by these technological developments our 3D shape recognition research focusses on reconstruction of 3D shape models from point clouds and on comparison of 3D shape models. Obtaining a 3D shape model by digitizing the surface of an object usually requires the acquisition of several 3D point sets from different viewpoints. As part of our contribution to the AIM@SHAPE network of excellence we work on the development of point cloud matching methodology, so that two point clouds can be stitched together automatically, without human interaction. The major problem in combining multiple point clouds is accurate registration, or matching, i.e. finding the transformation (translation, rotation) that relates part of one data set with part of another, such that a suitable distance measure between the two is minimized, while the matching parts are as large as possible. We want to solve the problems of finding the matching parts, finding the transformation without an initial approximate solution given, the large number of points (up to millions), and constructing a similarity measure that is usable for partial matching problems. Also, as part of our contribution to the AIM@SHAPE network of excellence we work in collaboration with Prof. Dr. Hein Daanen, TNO Human Factors in Soesterberg, the Netherlands, on adding semantics to human body models. From a database of human body models (provided by TNO Human Factors), we will derive anthropometric features, and identify meaningfull landmarks. In addition, we investigate the comparison of 3D shapes models by 3D shape matching which is an important ingredient in shape retrieval, recognition and classification, alignment and registration, and approximation and simplification. We have developed a new geometric approach to 3D shape comparison and retrieval for arbitrary objects described by 3D polyhedral models that may contain gaps. See for a description of this method and a demo the 3D shape retrieval engine webpage.