Intensity dependent spatial quantization with application in HEVC

With the increase of video resolutions used in multimedia applications, solutions for improved compression are sought. The next generation video compression standard, High Efficiency Video Coding (HEVC) is being developed by ITU-T and ISO/IEC MPEG with the goal to provide significant improvements over H.264/AVC. To further increase compression capabilities of HEVC, this paper proposes an additional bitrate saving mechanism. In this context, an Intensity Dependant Spatial Quantization (IDSQ) perceptual tool is proposed which exploits the intensity masking of the human visual system and perceptually adjusts quantization. The proposed IDSQ allows for adaptation to the video characteristics and its design meets low complexity implementation requirements. The proposed IDSQ has been tested in the HEVC test model and its performance is reported under “just noticeable distortion” or “perceptually lossless” conditions. In this setup, bitrate reductions of up to 25% are reported with an average of 3.4% over an exhaustive test setup.