Research On Motion Estimation And Error Control In H.264/AVC

With the extensive application of multimedia technology, compression coding technology for video is deeply developed. H.264 is a new generation video compression coding standard and is a hotspot in the current image communication field, which was established by the ITU-T Video Coding Experts Group and the ISO/IEC Moving Picture Experts Group. By adopting a number of new coding techniques, H.264/AVC has achieved a significant improvement in compression performance and a "network-friendly" video representation relative to existing standards.In this dissertation, we mainly research on the new techniques in H.264/AVC, emphatically on two core techniques: motion estimation and error control.Firstly, the new techniques in H.264/AVC are discussed; Experimental results prove the advantage of the new standard then. Secondly, we discuss the most importantmodule in H.264/AVC——motion estimation. Based on the new characteristics ofmotion estimation about H.264/AVC, we develop a new algorithm called a fast motion estimation method based on Structural Similarity (FMEBS). FMEBS introduces an image measurement called Structural Similarity to modify the distortion metric of Rate Distortion Optimized algorithm and uses the fast inter mode decision approach and efficient search patterns. Experiments show that FMEBS has advantages of both low delay and low bit error rate while obtaining good reconstruction image quality.Finally, the error control technologies that are relatively often adopted in video communication systems are explained in detail and the error control tools with the overall system analysis in the H.264/AVC standard are discussed in this paper. Two new error concealment algorithms are presented then. The area-directed adaptive intra error concealment algorithm uses the spatial correlation and texture characteristics of image to approximate the edges at the missing block and reconstruct the damaged area by interpolation. Experiments show that the edges at the missing block can be well recovered with the proposed algorithm and good results can be achieved for images of various content types. The temporal error concealment algorithm based on motion characteristics of the damaged area uses the spatial and temporal correlation of video to determine the correct mode of lost block and select the most appropriate motion vector to recover the damaged area among a set of candidate motion vectors. Experiments show that the new algorithm obtains good PSNR performance in the complicated motion region and good results can be achieved for video under various conditions of Packet Loss Rate.