Rate Control Algorithm for HEVC-SCC with Perceptual Features of Spatial and Temporal Gradient

Based on the large number of texts, graphics areas and sharp edges characteristics in screen content video, a rate control algorithm for HEVC-SCC with perceptual features of spatial and temporal gradient is proposed in this paper. Considering that effective bitrates should be used to the features that are more sensitive for human visual system, the perceptual features are extracted from spatial and temporal domains. In spatial domain, a difference operator is used to extract the gradient magnitude information and obtain the descriptor of spatial features. In temporal domain, gradient similarity measure information is used, which can better describe the perceptual features of screen content video. Combining the correlation between adjacent frames, the descriptor of time domain features is obtained. The fusion of these features is used to guide the rate control of HEVC-SCC. Experimental results show that, in random access mode, compared with the original HEVC-SCC, the perceptual qualities of videos are improved with the reduction of bitrate error.