Title :
Directional de-blocking filter for blocking artifacts reduction
Author :
Pasiyawala, Pratik ; Shah, Neil ; Modi, Meghavi
Author_Institution :
Electron. & Commun. Dept., Sarvajanik Coll. of Eng. & Tech. Surat, Surat, India
Abstract :
Block based motion estimation and motion compensation techniques are widely used in most of video coding standards including H.264/AVC (Advance Video Coding) to compress raw video. Motion compensated block is transform coded and quantize before entropy coding to achieve high coding efficiency. All such block based operations introduces annoying blocking artifacts and degrades the quality of reconstructed video. In order to improve the quality of the reconstructed video, several de-blocking algorithms have been proposed, which first classify blocks based on activity before applying appropriate filter. This paper proficiently classifies blocks according to horizontal and vertical activities as smooth, intermediate and complex blocks and applies directional deblocking algorithm to motion compensated frames to remove artifacts efficiently. Experimental results indicate that directional de-blocking algorithms outperforms for highly compressed standard video sequences due to appropriate block classifications.
Keywords :
data compression; entropy codes; filtering theory; image classification; image reconstruction; motion compensation; motion estimation; video coding; H.264/AVC; advanced video coding; artifact removal; block based motion estimation; block classifications; blocking artifact reduction; compressed standard video sequences; directional deblocking filter algorithm; entropy coding; motion compensated frames; raw video compression; reconstructed video quality; video coding standards; Classification algorithms; Filtering; Filtering algorithms; PSNR; Standards; Video sequences; Blocking artifact; De-blocking Block (DB); Macroblock (MB); Post filter;
Conference_Titel :
Engineering (NUiCONE), 2013 Nirma University International Conference on
Conference_Location :
Ahmedabad
Print_ISBN :
978-1-4799-0726-7
DOI :
10.1109/NUiCONE.2013.6780124