Adaptive coding configuration for multi-resolution video coding

Abstract

In this paper, we present the efficiency analysis of two sophisticated coding tools, multi-reference frame (MRF) and variable block size (VBS) that cost high computational complexity, in various spatial resolutions. The relationship between coding efficiency and spatial resolution is theoretically discovered. Based on the conclusion of the efficiency analysis that the efficiency improvement from sophisticated coding tools are gradually decreased in higher resolution, we propose an adaptive coding configuration for the encoding with various resolutions that yield significant complexity reduction with negligible RD performance decrease.