DocumentCode
698219
Title
Improved context-adaptive arithmetic coding in H.264/AVC
Author
Karwowski, Damian ; Domanski, Marek
Author_Institution
Multimedia Telecommun. & Microelectron., Poznan Univ. of Technol., Poznan, Poland
fYear
2009
fDate
24-28 Aug. 2009
Firstpage
2216
Lastpage
2220
Abstract
An improved Context-based Adaptive Binary Arithmetic Coding (CABAC) is proposed for applications in video compression. The improved CABAC is characterized by higher coding efficiency as compared to that of the original CABAC as standardized in ISO MPEG-4 AVC (Advanced Video Coding) and ITU-T H.264. In the proposal, the increase of compression ratio is obtained by application of more exact data statistics estimation based on well-known data modeling algorithm: Context-Tree Weighting (CTW). The improved CABAC codec has been tested in the framework of AVC video codec. The extensive experiments show that application of the modified AVC yields 1.2% - 5.4% bitrate reduction. Computational effort of the modified AVC decoder is about 5-10% and 15-25% greater relative to the original AVC decoder for SD and HD video sequences respectively.
Keywords
adaptive codes; arithmetic codes; binary codes; data compression; image sequences; statistical analysis; video codecs; video coding; AVC decoder; AVC video codec; CABAC codec; CTW; H.264/AVC; HD video sequences; ISO MPEG-4 AVC; ITU-T H.264; SD video sequences; advanced video coding; bitrate reduction; coding efficiency; compression ratio; context-based adaptive binary arithmetic coding; context-tree weighting; data modeling algorithm; data statistics estimation; video compression; Abstracts; Automata; Cities and towns; Encoding; Engines; Indexes; Quantization (signal);
fLanguage
English
Publisher
ieee
Conference_Titel
Signal Processing Conference, 2009 17th European
Conference_Location
Glasgow
Print_ISBN
978-161-7388-76-7
Type
conf
Filename
7077794
Link To Document