Search

A study on the use of post-processing in video compression proposes a novel, postprocessing algorithm to enhance the visual quality of low bit-rate, MC-block-discrete cosine transform-based image and video compression systems. The algorithm minimizes the unwanted blocking effects and mosquito noise artifacts while maintaining image sharpness and naturalness. It is relatively simple yet performs quite well in different situations.

A postprocessing algorithm that enhances video picture quality adapts itself to pixel characteristics in a video sequence to filter out noise and preserve image detail. The system puts video signals into different classes and uses separate nonlinear filters matched to each class. The most general version of the algorithm uses motion-compensated frame averaging to improve picture quality. A suboptimal, spatial version of the adaptive postfilter enhances picture quality in a 384 kbps 3-D subband coder.

A three-dimensional spatio-temporal subband decomposition method developed for video coding functions by arranging the image sequence into various spatio-temporal frequency bands and then separating the information into spatio-temporal subbands. The method yields a 1-Mbps coder using an adaptive differential pulse code modulation scheme and adaptive bit allocation. A new vector quantization scheme is suitable for data in higher frequency bands.