Image multiple description coding method based on interleaving extraction and block compressive sensing strategy

Based on Interleaving Extraction and Block Compressive Sensing (IEBCS), a new Multiple Description Coding method (IEBCS-MDC) which can be achieved real-timely during imaging process is presented. The method is first partitions an image into several sub-images using interleaving extraction, then measures each sub-image with block compressive sensing and forms multiple descriptions. At the decoding terminal, the method reconstructs the original image by solving an optimization problem. Block strategy ensures that the complexity of measurement process does not change due to image size, so the method is simple and easy to implement, suitable for handling high-resolution images, and the characteristic self-recovery capability enhances the ability against packet loss. Experimental results show that, compared to CS-MDC, the proposed method can handle much bigger images in the same hardware environment and the reconstruction quality is also better than CS-MDC with the same packet loss probability.