Abstract
In this thesis, we propose an automatic video conferencing system for background substitution. Since humans are the principal subject in these videos, our framework is based on human shape clues to separate humans from complex background and replace or blur the background for immersive communication. We first detect face position and size, track human boundary across frames, and propagate the segmentation likeihood to the next frame for obtaining the trimap to be used as input to the Random Walk algorithm. Besides, we also include gradient magnitude in edge weight to enhance the Random Walk segmentation results. In this part, we demonstrate the effectiveness of the proposed background substitution system through experiments on some real videos. We also present a system based on a multi-core processing architecture. Two tables, TYPE and INDEX, are introduced to fast locate the required data for the close-form solution. We demonstrate the parallelization strategies for the proposed fast RW algorithm and face detection on heterogeneous multi-core embedded platform to make the most use of the system architecture. Compared to the single processor implementation, the experimental results show significant speedup of the parallelized human background substitution system on a multi-core embedded platform, which consists of an ARM processor and two DSP cores.