Research Of Screen Content Coding Based On HEVC Standard

With the surging development of Internet information technology and video streaming technology,720p and 1080p HD resolution video becomes increasing popular. The emergence of 2K resolution and 4K Ultra HD resolution video causes the explosive growth of video data amount, which is the challenging problem to big data network transmission and storage. To high-resolution and ultra-high-resolution video, the effective compression under the condition of no quality loss is one of the hot topic in current research. In addition, with the popularity of mobile devices, remote desktop, virtual desktop and wireless display applications, how to reduce the encoding complexity and the bandwidth required in the network transmission process according to the specific content in the certain scenes is the urgent problem. High efficiency video coding (HEVC) mainly focuses on the natural capture video sequences, and has achieved effective compression under a certain degree of complexity. Comparing to natural capture video, screen content video shows difference on the source of video format. Aiming at screen content video, the novel screen content coding (SCC) is developed on the base of traditional HEVC coding standards.Comparing to natural capture video, screen content video consists of more high-frequency and low-correlation content. According to the characteristics of the screen content of the image, the screen content encoding on the basis of HEVC added some new encoding tools, such as intra block copy, palette mode, adaptive color transformation and adaptive motion vector resolution.This paper mainly introduces a coding framework of high efficiency video coding standard, the quadtree coding structure and predictive coding technique. Then we describes the new tools of screen content video on this basis, such as intra block copy, palette mode, adaptive color transformation and adaptive motion vector resolution.For the input source of different sample formats in the screen content video, we take advantage of video quality assessment method to study its effect on the screen content video compression efficiency. As seen, we can trade 1.20% and 1.36% SSIM loss for 28.79% and 33.05% bit rate reduction of respective AI and LB configurations when we use YUV444 instead of RGB signal for screen content coding. Furthermore, we have 3.56% and 3.41% SSIM loss for 31.74% and 34.49% bit rate reduction of AI and LB configurations respectively when feeding the YUV420 signal instead of RGB into screen encoder.For highly complex issue of screen content encoding, we propose the fast intra coding optimization algorithm, including fast intra mode decision based on background detection and fast block matching based on adaptive searching step size adjustment. Compared with the conventional method, this method greatly improves the coding performance. In total, the proposed solution achieves 39% encoding time reduction and 0.7% BD-rate increase for AI lossy case, and 35% encoding time reduction and 0.2% bitrate increase for AI lossless case respectively.