Reconstruction of facial images by using facial motion and image processing techniques

Abstract

The facial expressions of an human being can be explained by the motion of the facial features. If we are able to define these expressions by some motion vectors it will be possible to reconstruct the original face image from the face image with expression. In this context, the studies of this paper are focussed around the motion vector estimation and image reconstruction. For the motion vector estimation two well-known block matching techniques, namely Full Search Algorithm and Three Step Search Algorithm, have been used and the results have been compared via PSNRs (Peak-to-Peak Signal-to-Noise Ratio) and time consumptions.