Compressed sensing (CS), due to its superior performances that very few amounts of coefficients are capable of reconstructing the image with reasonable quality, is an effective means in data compression in recent years. For the transmission of image compressed with CS, due to the channel errors that may be experienced, data protection would be required. We employ multiple description coding (MDC) for protecting color images with compressed sensing. With MDC, compressed information can be divided into different chunks at the encoder, and all the chunks can be transmitted over multiple independent channels to alleviate the influences caused by channel errors. At the decoder, the possibly corrupted chunks are assembled together, and MDC can be applied to reconstruct the image. With our algorithm, simulation results presented the error resilient capability and improved quality of color images.
2015 Third International Conference on Robot, Vision and Signal Processing (RVSP), pp. 63-66