Real-time low-bitrate multimedia communication for smart spaces and wireless sensor networks

摘要

Smart spaces is a scenario that is ideal for the application of wireless communication technology to benefit non-expert users. A prime example of such an application would be low-cost and real-time video surveillance of hallways or video conferencing. Traditional solutions for these scenarios use state-of-the-art video codecs like H.264/AVC in order to achieve the low bitrate which is essential in wireless communication. On the flip side, such encoders are computationally complex and need expensive hardware to run on thus making them inefficient choices for being deployed in clusters. Recent research has resulted in distributed video coding (DVC) being proposed as a solution for applications that have limited battery resources and low hardware complexity, thus necessitating a low-complexity encoder. It is now a popular topic in the research community and the past years have seen several implementations. However, current DVC solutions use iteratively decodable channel codes like low-density parity check - accumulate (LDPCA) codes or Turbo codes that have large latencies. In order to make real-time communication possible, the proposed architecture makes efficient use of skip blocks to reduce the bitrate, eliminates the iterative decoding nature of the Wyner-Ziv (WZ) channel and uses a simple data-hidingbased compression algorithm. This drastically cuts down on the time complexity of the decoding procedure while still maintaining a rate??distortion performance better than that of H.264/AVC intra-coding and other current DVC solutions.