Extremely Randomised Trees for Computational Complexity Reduction of Omnidirectional Intra Video Coding
Filipe, J.
; Carreira, J. F. M.
;
Távora, L.M.
;
Faria, S.M.M.
;
Navarro, A.
;
Assunção, P.A.
Extremely Randomised Trees for Computational Complexity Reduction of Omnidirectional Intra Video Coding, Proc Portuguese Conf. on Pattern Recognition - RecPad, Évora, Portugal, Vol. , pp. 77 - 78, October, 2020.
Digital Object Identifier:
Download Full text PDF ( 205 KBs)
Abstract
This paper presents a novel method to reduce the computational complexity of intra-coded 360º video in Equirectangular Projection (ERP) format. The proposed method is based on three Extremely Randomised Trees models to predict the maximum partition depth that should be used for intra-coding of the complex nested data structures (Quad Tree, Binary Tree and Ternary Tree) used in the forthcoming video coding standard (Versatile Video Coding). The results show that an average complexity reduction of 31.35% is achieved, with a negligible loss of coding efficiency of just 0.42%, outperforming other state of the art low complexity solutions.
