MSc Semester Project

Title

Multiview video coding

Candidate

Philippe Hanhart

Supervisor

Prof. Dr. Touradj Ebrahimi

Assistant: Dr. Lutz Goldmann

Expected

January 8, 2011

Place

EPFL

Description

Most video processing and coding systems rely on one single camera, referred to as monoview approach. In the last two decades, extensions to two-camera solutions (also referred to as stereo) have been investigated with limited success in both coding and video analysis applications. Although multiview is also used in solutions with two cameras, in general the term is only used for solutions that use more than two cameras. Multiview video processing has attracted increasing attention recently and has become one of the potential avenues in future video systems, thanks to the reducing cost of cameras. Many tasks can benefit from the availability of multiple views of the same scene, such as interpolation, restoration, segmentation, object recognition, etc. On the other hand, the amount of data captured in multi-view systems is often tremendous, which makes the coding efficiency critical for such systems to make them practical.

MPEG saw the importance of multiview video coding and has been working on the exploration of 3DAV (3D Audio-Visual) technology. Our interest is mainly in multiview video coding research perused in 3DAV in which several companies involved in MPEG’s research showed a lot of interest lately since the use of multiview video systems is increasing in different fields such as Cinematographic Special Effects, Video Surveillance, Virtual Environments Generation, etc.

This project aims at the development and implementation of a view prediction tool, across different views, which performs at least as good as the state of the art. Then, Implement it in the MPEG Multiview software.

The student will try to follow the following plan: