Image based face recognition systems: new perspectives

Research Area:
Research Lines:
Research Lines:
Responsible:
Research
-
Start date: 2004-01-01
Length: 22 months
Project abstract
The project was oriented toward the study and the development of new techniques for the coding of moving images (image sequences). In particular scalable video coding has been considered, moreover techniques based on the DVC (Distributed Video Coding) approach has been investigated. Scalable video coding aims to realize a coded stream from which different sub-streams characterized by different bit-streams cane be extracted with out deconding-and-recoding operations. It is extremely useful for audio-video distribution over heterogeneous networks. Instead, the DVC paradigm is based on the idea of moving computational complexity from the encoder device to the network or to the decoder engine. The project is closed; a patent relative to the use of the DVC technique has been filed.
Project results
The project, ended in October 2005, has as results the realization of video coder simulator that is able to generate scalable streams and uses a DVC approach. Several contribution have been published in high level international congresses on image/video processing, in particular for what regards the use of the DVC techniques to create robust coded streams to be transmitted over channel characterized by a high loss rate (wireless channels).