Comparisons of Reconstruction Capabilities for Lossy Transmission with Block-Based Compressed Sensing

Abstract

Robust transmission of compressed signals is much required in multimedia communications. Compression of multimedia is inevitable, and we choose block-based compressed sensing to play the role for compression. Compressed coefficients are transmitted over lossy channels. Channel errors or losses may cause degradation to reconstructed image qualities after reception. We consider making use of different reconstruction methods at the decoder and make comparisons. In addition, error resilient techniques may also help to enhance reconstruction quality. Multiple description coding can be applied to work with different reconstruction schemes, and enhanced reconstructed image quality may be expected with the integration of error resilient techniques. Simulation results have shown the capability of reconstruction schemes, and the advantages for error resilience to obtain enhanced quality.