AN IMPLEMENTATION OF MPEG IN MULTI-TERMINAL VIDEO COMPRESSION.pdf

Abstract

Video sensor networks (VSNe) are basically application related setups. These platforms have their own requirements and constraints. But, all these applications have one common problem of bandwidth requirement. Each node within a sensor network has to transmit its finding to a central receiver. Each terminal's information is important and carries a significant part of final results that must be transmitted to get best results. Bandwidth of the link is limited and it is a great problem to transmit all nodes findings within this available link capacity. Multimedia applications have huge textual as well as visual data. They, mostly comprise of repetitive patterns. Current video coding standards eliminate such data redundancy by exploiting spatial redundancy (within a video frame) as well as temporal redundancy (among video frames). These video codec standards are deployed on each sensor node that does not save the bandwidth requirement up to an optimum limit, since it does not exploit inter-sensor redundancy. Same video codec standards like MPEG can be deployed in a multi-terminal network by extending the transform coding of motion compensated coefficients among different sensor at a regular pattern. Deploying a standard video codec in this fashion exploits the inter-sensor redundancy, thus yields a great saving in required bandwidth for the resultant piece of information coming out of a wireless sensor network. This research work is an extension of MPEG, a video compression standard for a multi-camera setup. Basically, in this research study camera sensor nodes of a wireless network are allowed to communicate in two strategies, strategy one allows minimum communication. While sensor nodes are allowed to communicate a bit more in strategy two. In both communication strategies, system has been designed to work in two working-modes: named as Scenario A and Scenario-B. Scenario A takes more processing time but gives large PSNR values while system in Scenario B can be deployed in a situation where quick system response is required provided degraded video quality can be compromised. Results in tabular as well as graphical forms have been formulated to evaluate the system performance in both working modes under both communication strategies. Overall system shows optimum performance at low bitrates.