Wireless Video Anti-error Transmission Technology

Following with great improvement of technologies on telecommunications and video processing, digital video can be used in more and more applications now. However, source video has a huge mount of information, and makes it hard to be stored or transmitted over networks. With the help of successive feature both in spatial and temporal fields, we can remove residual information existing in video data. High compression ratio is one target in the process of video codec standards. And now H.264/AVC does better comparing with other standards. But highly compressed video data means being vulnerable to errors. Once errors occur in a frame, they will disperse in the frame, and even affect following frames seriously. In the case of wireless networks where video signals are exposed in the air and errors occur easily, task on resilience is much more important when transmitting video.In this paper, we will analyze reasons for causing errors, and then describe new features of error resilience in H.264/AVC codec standard. Our researches on resilient transmission are mainly in two aspects, one from video decoder while the other from video encoder. At decoder side, we focus on how to apply error concealment technology to recover damaged video, and try to improve video quality as much as possible. After analyzing previous algorithms, we will propose two improved error concealment methods based on region segmentation. One of them modifies the method on edge point searching. The other alters algorithms for edge detecting and broken edge recovering. At encoder side, we will study how to adopt multiple description coding (MDC) technology to reduce error probability, and make sure decoding continues even when errors occur. By researching the way producing multiple descriptions, we will propose a new MDC method based on altered big and small frames.We have implemented these proposed algorithms on JM15.0 platform of H.264/AVC standard, and taken some tests and analysis. Comparing with the initial algorithm, two proposed error concealment algorithms can effectively improve subjective and objective video quality. Altered big and small frames based MDC method can enhance video quality, comparing with the conventional MDC algorithm and the standard single description coding method at the approximate bit rate.