HEVC-based adaptive quantization for screen content by detecting low contrast edge regions

Hong Zhang, Oscar C. Au, Yongfang Shi, Xingyu Zhang, Ketan Tang, Yuanfang Guo

Abstract: High-Efficiency Video Coding (HEVC) is the newest video coding standard which can significantly reduce the bit rate by 50% compared with existing standards. The key features and new tools in HEVC are designed for natural video sequences captured by a real camera. Different from natural videos, screen content contain much more edges in text and icon regions. The current video coding standards may blur or even remove low contrast edges, which are very important in screen content for human eyes to recognize the character and the icon. Therefore, this paper proposes an effective modification on HEVC to preserve the low contrast edges in screen content. First, discrete laplacian filter is adopted for edge detection, and then we adaptively adjust QPs for low contrast edge regions, which can be detected based on our designed measurement for edge contrast. Experimental results show that nearly all the regions containing low contrast edges can be detected, and the adjustment of QPs for these regions can greatly protect the edges with no RD performance reduction.