This contribution describes the creation of highly realistic 3D models of participants for distributed 3D videoconferencing systems. These models consist of a flexible triangular mesh surrounding an interior skeleton structure, which is based on a simplified human skeleton. The vertices of the predefined mesh template are arranged in rigid rings along the bones of the skeleton. Using 3D data obtained by a stereoscopic approach the size and shape of this initial mesh is adapted to the real person. Adaptation allows to texture the model from real images leading to a natural impression. The mesh organization in rigid rings gives an efficient way for surface deformation according to the skeleton movements. The resulting model is transmitted once and subsequently animated using the simple parameter set of the interior skeleton structure.