Perceptual video quality optimization in AWGN channel using low complexity channel code rate allocation

In error-prone channels, forward error correction is necessary for protecting important data. In this paper, we use a packet loss visibility model to evaluate the visual importance of video packets to be transmitted. With the loss visibility of each packet, we use the Branch and Bound method to optimally allocate rates of Rate-Compatible Punctured Convolutional codes. The complexity of our prior algorithm can be reduced by k-means clustering before using the Branch and Bound method. Experimental results show that the proposed unequal error protection algorithm can improve upon the received video quality compared to our prior work with much lower complexity.