Study On Enahnced Technology Of Video Coding

Recently, digital video has been widely applied in areas range from digital television, video conferencing, video telephony, internet streaming to multimedia storage. With wide application of computer and network technology, video compression technology has been developing rapidly. Continued development of video compression technology has been promoting wider and wider application of digital video. Therefore, it is of great significance to study on video compression technology.The main objective of video compression is to achieve video compression ratio as high as possible. Video compression technology has made breakthrough development in the last 20 years. The latest video coding standard H.264/AVC provides much higher coding efficiency than the early video coding standard (e.g. MPEG-1). At present, H.264/AVC is widely adopted as video coding tools by a large number of companies in the world and applied in various areas of digital video. However, H.264/AVC can't always guarantee its high coding efficiency to all video sequences because of their complicated statistical features. Many video preprocessing technology (e.g. pre-filtering) and special coding methods (e.g. texture analysis) are introduced to achieve higher compression ratio. We call these special coding methods - enhanced technology of video encoding for short.This paper studies on enhanced technology of video coding in H.264/AVC standard. The work aims to present some enhanced coding methods to improve coding efficiency without additional coding complexity. Serveral enhanced coding methods are presented to improve H.264's coding efficency in this paper as follow:First, we improve some intra prediction methods to in H.264 standard. In the paper, we give a brief introduction to intra prediction modes in H.264; analyze their shortcomings; and present corresponding improved intra prediction based on recursive method, which predicts pixels more accurately. Then, we improve intra prediction based on BMA method, which is initially presented by Peking University. At last, we apply intra prediction based on recursive method and BMA method together to intra coding in H.264/AVC standard. Experiment shows that our algorithms can achive higher coding efficiency for all test sequences; outperform the performance of H.264/AVC's algorithm by average 0.27dB.Second, we improve lossless intra coding in H.264/AVC standard. In the paper, we improve lossless intra coding in H.264/AVC by introducing down sampling method. Experiment shows that average 5%-10% bit-rate can be reduced. Then, we present interlaved prediction based on down sampling method for better prediction quality. Experiment shows that average 9%~15% bit-rate can saved.At last, we present a highly efficient scheme for encoding panoramic video, which is H.264 compatible. We don't encode panoramic video directly. In the new scheme, we convert the original video to a panoramic image, and then the panoramic image to standard H.264/AVC bit-stream indirectly. Experiment shows that bit-stream encoded with the new scheme is only 1/10 of bit-stream encoded with H.264/AVC standard directly.