Multiple Descriptions and Adaptive Delivery of Media Content over Wireless

The dynamic and varying wireless channel makes it a challenge to efficiently delivery media contents over wireless with quality guarantee. We propose a multi-description scheme for the adaptive delivery of media contents over wireless. A video segment is abstracted at different levels, such as syntactic and semantic description by feature extraction and keywords annotation, video abstract by key frames extraction, transcoded video (or video objects) by rate adaptation, etc., thus multiple descriptions of a video segment can adapt to all wireless scenarios. Furthermore, these descriptions can be stored on servers or wireless nodes; it can be exchanged efficiently between nodes and servers, nodes and nodes. We give an example to illustrate the adaptation process under different constraints.