An Attention-Based Network-Aware Adaptive Intra Refresh Method for Wireless Video Communications

A novel adaptive intra refresh method for the wireless video communication is proposed, which considers the characteristic of the human visual system and the state information of the time-varying wireless channel jointly. An attention model is used to compute the perceptual weight map of the video content. The dynamic feedback channel state information (CSI) of the wireless channel, such as bit error rates and packet lost rates, is utilized to predict the channel distortion. Experimental results indicate that the proposed method significantly outperforms other methods without considering the CSI.