An innovative knowledge-based scheme is presented for videophone sequence coding, with a reasonable subjective reconstruction quality. The videophone images are segmented and approximated by means of a suitably adapted 3D wire-frame model capable of faithfully reproducing the signal temporal evolution through a limited number of updating parameters extracted from the real sequence. The muscle modeling represents the basic issue for the analysis of facial expressions at the encoder and for their synthesis at the decoder. Facial muscles are modeled either as linear fibers or as circular fibers and are characterized by descriptive parameters such as thickness, length, position and direction. The animation is performed by simulating the muscle contraction and relaxation over the domain of influence defined by the muscle parameters. The interactive procedure for the construction and adaptation of the 3D model and the analysis/synthesis algorithms, implemented on a multiprocessor AT&T Pixel Machine, are described, together with some examples of the achieved experimental results

Deals with a robust technique which is able to synthesize the volumetric structure of an object from a set of 2D perspective views where the object silhouette has been extracted. The surface texture is also recovered from the views. The final model is a closed polygon mesh surface integrated with texture informations, and it is thus suitable for standard graphic rendering. The proposed procedure can be used to obtain real world object representations starting from real views instead of manually building them by means of solid modeling and painting systems. As shown by the reported experiments, the results are promising in all cases where the objects can be accurately reconstructed from the occluding contours technique, i.e. when the concavities, if present, are visible in the projected object silhouettes.<>

A regularization procedure for the volumetric reconstruction of objects from occluding contours is presented. It exploits a simple geometrical constraint on the intersection of the volumes generated by backprojecting the object silhouettes and minimizes the volume reconstruction error with respect to the extrinsic parameters, i.e. position and orientation, of each view. The results of simulation tests are described, showing very good improvements in the reconstruction quality.

Keywords: LTS1 Reference LTS-CONF-1991-016 Record created on 2006-06-14, modified on 2016-08-08