Reseach On Image Error Concealment Algorithm Based On MV Spatial Correlation

Since multimedia technology is developing rapidly nowadays, people’s demand of video communication is becoming higher and higher, so the traditional video compression technology cannot satisfy the demand of the public. In order to provide better quality of video communication and larger compression ratio, H.264/AVC codec comes out, which provides up to 50% decrease in bit-rate compared with previous standard, with no decrease in video quality. This greatly reduces the video transmission pressure on the network channel.However, in actual communication, signal loss often happens in transmission since the channel is unstable, resulting in signal distortion in terminal. For image and video signal, this will result in black block, Mosaic and even the loss of the whole image. The so-called error concealment is a method to estimate the lost content by using the correlation of the accurate information in the decoder side, in order to restore the image quality. Such technology will improve the video quality and won’t bring any pressure to the transmission channel, so it gets the favor of image transmission system and the related standard.In this paper, two fast and efficient error concealment algorithms based on the MVs’ correaltion are proposed. And the main works have been done are listed below:1. Firstly a brief introduction to the principle of multimedia video coding and the history of coding standard is given. The description focuses on the workflow of coding standard, some new technologies in H.264/AVC, and its application in the daily life at present.2. Some researches in the error control of H.264/AVC have been done in the paper, and these technics are divided into two kinds: encoding side(forward) and decoding side(backward). This paper emphasizes on the temporal error concealment technology.3. Several kinds of temporal error concealment algorithms based on H.264/AVC in the latest years have been studied. The paper describes them in detail and lists their advantages and disadvantages respectively.4. In this paper, an effective error concealment algorithm MVSC is proposed, which utilizes temporal information(MV) and the spatial correlation in the same frame to recover the lost MB. Simulation results show that the proposed algorithm can achieve better subjective and objective qualities compared with the existing temporal EC algorithms, and can be applied into practical use.5. Some improvements in the MVSC algorithm are made and the parameters of motion trend are redesigned. Then we propose a new method call OMVSC. From the simulation we can find the results show satisfied improvement compared with MVSC.