Scalable, subband-based video coding with a locally-adaptive perceptual distortion measure

Existing video coders do not fully exploit the local variation of perceptual masking thresholds. This paper presents a locally-adaptive subband-based video coding scheme which is based on a perceptual distortion measure. The local adaption does not require any additional side information since the perceptual thresholds are computed based on the already transmitted data. The proposed coding scheme is scalable in frame rate and spatial resolution. It supports a fixed rate and perceptual target distortion modes. It can be easily generalized to DCT-based methods.