Design And Implementation Of HEVC Intra-prediction Algorithm Based On Pixel Gradients

With the development of Internet technology, more and more people prefer to watch online high definition(HD) video using smartphones and tablet PCs. Meanwhile, Ultra HD(UHD) technology is growing up gradually. Online HD video causes heavy burden to network devices, and the meanstream video coding standard, H.264, has a lower compression performance on HD and UHD video. To further reduce the coding rate and improve compression performance on HD and UHD video, JCT-VC has developed a new coding standard HEVC.HEVC is based on the block-based hybrid conding architure, combining flexible division of coding units, more precise intra-frame prediction, in-loop deblocking filtering and so on. But new technology brings much more computation to intra-frame prediction of HEVC. Therefore, how to effectively reduce the intra-frame coding time has been a research hotspot.We improve the intra-frame prediction algorithm by using the gradient information of pixels, and evaluating the performance of the algorithm with the use of C++ language. In the end, we analyze the results of the performance as follows:Firstly, we analyze the basic coding structure and the key technology of intra-frame prediction. And we compare it with corresponding in H.264.Secondly, we analyze the intra prediction algorithm of H.264 and HEVC. Based on the analyses, we improve the intra-frame prediction algorithm by using the gradient information of pixels, and evaluate the algorithm using C++ language.In the end, we compare the performance under the all-intra and low complexity condition. Accroding the evaluation, we find that the proposed algorithm can reduce the total coding time by 22.50% with slightly 1.62% BD-rate increase.We improve the intra-prediction algorithm using the gradient information of pixels and reduce the total coding time effectively. It is of certain theoretical value and practical value.