Adaptive Error Concealment Technology For Stereoscopic Video With Low Complexity

With the rapid development of video technology, stereo video caused widespread concern of researchers for its vivid visual experience.The application of stereo video in various fields has become increasingly rich and colorful. Stereo video data transmission on the network increases exponentially, due to noise, bandwidth, delay jitter and other causes of instability can lead to data loss and the frame decoding error. Resulting in the decline in the quality of video, can’t even normal display, and it eventually affects watch user’s visual effects. In order to solve the problem, this paper adopts an adaptive error concealment algorithm. In this dissertation, specific studies are as follows:First of all, this paper proposes a transmission distortion estimation model for stereoscopic video with specific error concealment method. Considering intra-view propagation distortion, inter-view propagation distortion and the distortion caused by error concealment, according to different network conditions, using the method of low complexity, the model can estimate accurately end-to-end transmission distortion with low complexity. The experimental results show that, for different packet loss rate, the average prediction error of the distortion model can be controlled within 6%. Considering different network environment, the model can be adapted to estimating transmission distortion for stereo video sequences with different resolution.Secondly, by using the low complexity stereo video coding structure, IPPP, in order to shorten the encoding and decoding time. As the lost macro blocks exist in stereo video, we can adopt error concealment algorithm relying on the temporal correlation between adjacent frames and the spatial correlation between viewpoints. According to the motion characteristics of the macro block in current frame, we can estimate the motion characteristics of the lost frame. For frames consist of faster moving objects, the reconstruction block was screened in particular candidate blocks using the improved outer boundary matching algorithm. Select the candidate blocks higher priority weighted to get the new candidate block, and then use the new candidate block recovery of damaged macro block. Experiments show that the algorithm in this thesis, without increasing computational complexity, can effectively improve the quality of restored stereo video sequences with different quantitative parameters and different packet loss rate.Finally, by analyzing the scene change in stereo video frame, an adaptive error concealment algorithm for different scenes is proposed to improve the image quality. Combining with frame difference method and threshold method, this paper proposes a new method to detect the position of the scene change frame. We decide the motion state of the whole frame according to the macro block motion state for each frame. According to the different motion state of stereo video to choose the corresponding error concealment method, the paper put forward an adaptive error concealment algorithm for stereo video with low complexity.The experiments based on JM show that considering effect and decoding complexity, the adaptive error concealment algorithm has better effect on processing of macro block loss recovery in stereoscopic video image.