Computationally efficient mode selection in H.264/AVC video coding

Abstract

In the H.264/AVC video coding standard, the mode selection part requires a considerable amount of computation and memory access. This paper presents a fast or computationally efficient mode prediction and selection approach which has the following attributes: (a) both the spatial and temporal information are used to achieve early termination using adaptive thresholds, (b) inclusion of a modulator capable of trading off computational efficiency and accuracy and (c) a homogenous region detection procedure for 8times8 blocks based on adaptive thresholds. It is shown that the above attributes lead to a computationally efficient video coding implementation as compared to the previous fast approaches. The results obtained on QCIF, CIF, and HD format video sequences based on two different implementation platforms JM12 and x264 are presented to show the computational efficiency of the developed approach at the expense of acceptably low losses in video quality. Improvement in memory access performance is also demonstrated using a DSP processor simulator.