Model-BasedErrorConcealmentfor.PDFVIP

Model-Based Error Concealment for Wireless Video Deepak S. Turaga and Tsuhan Chen Electrical and Computer Engineering Carnegie Mellon University, Pittsburgh, PA 15213 dturaga@alumni.cmu.edu, tsuhan@cmu.edu Abstract We introduce model-based schemes for error concealment of networked video. We build appearance models for specific objects in the scene and use these models to replenish any lost information. Due to the models being designed specific to the object, they are able to capture the statistical variations in the object appearance more effectively, thereby leading to better error concealment performance. We examine statistical modeling techniques from literature and introduce a new efficient and accurate linear model for data representation called the Mixture of Principal Components (MPC), and use these models for error concealment. We simulate lossy network conditions and show that these model-based concealment schemes outperform the traditional concealment schemes across a variety of loss probabilities and bit rates for the coded video. 1. Introduction Loss of data during transmission of compressed video leads to objectionable visual distortion in the decoded video. In order to minimize distortion at the decoder end, many schemes for error resilience and error concealment have been developed. Error resilience includes schemes at the encoder end where some redundancy is introduced in the bitstream that makes it possible to recover lost information. Examples of these include error correcting codes, data partitioning techniques etc. Error concealment involves post-processing of the video at the decoder end to hide the effect of the transmission errors. Both error resilience and error concealment techniques may be used in conjunction to improve the quality of the decoded video. The focus of this paper is on error concealment t