Edge adaptive graph-based transforms: Comparison of step/ramp edge models for video compression
Yung‐Hsuan Chao, Hilmi E. Egilmez, Antonio Ortega, Sehoon Yea, Bumshik Lee · 2016
In this paper, we propose a new edge model for edge adaptive graph-based transforms (EA-GBTs) in video compression. In particular, we consider step and ramp edge models to design graphs used for defining transforms, and compare their performance on coding intra and inter predicted residual blocks. In order to reduce the signaling overhead of block-adaptive coding, a new edge coding method is introduced for the ramp model. Our experimental results show that the proposed methods outperform classical DCT-based encoding and that ramp edge models provide better performance than step edge models for intra predicted residuals.