Error Resilient Techniques for Wavelet Based Compressed Sensing

Compressed sensing (CS) is a newly developed research in multimedia compression studies. A large portion of papers in this field focuses on compression performance. In this paper, we aim at transmitting CS signals over lossy channels. We employ wavelet transform, and then apply compressed sensing to original image to obtain CS signals. Channel errors may cause severe effects to CS signals, leading to degradation to the reconstruction. We use multiple description coding (MDC) to protect CS signals beforehand, transmit the protected signals, and recover the reconstruction with the calculations from the compensations of MDC. Simulations have presented the alleviation of reconstructed image quality with proposed algorithm.