An H.264/AVC inverse transform adaptation method for video streaming applications

Bandwidth sharing in video streaming applications becomes problematic with the explosion of live services. In such applications, compression optimization techniques are seen as an efficient way for enforcing high quality video transmission. In this paper, we propose a low-complexity adaptive encoding framework for H.264/AVC, in which the inverse transformation matrix is optimized for each frame, as function of both content and quantization noise. As the transform optimization is done in one encoding pass, it can be successfully used in live encoding setups. The proposed synthesis filter adaptation method shows promising results compared to H.264/AVC, making it suitable for video streaming applications.