DocumentCode :
514439
Title :
LA-RDO based error resilient coding using multi-layer Lagrange multiplier selection in SVC
Author :
Shoaib, Muhammad ; Cai, Anni
Author_Institution :
Multimedia Commun. Centre, Beijing Univ. of Posts & Telecommun., Beijing, China
Volume :
2
fYear :
2010
fDate :
7-10 Feb. 2010
Firstpage :
1600
Lastpage :
1605
Abstract :
Loss-aware rate distortion optimized (LA-RDO) MB mode decision has been introduced in scalable video coding (SVC) to minimize the quality degradation impact of transmission errors. It aims to optimize rate-distortion performance considering transmission error and resulted error propagation. However, Lagrange multiplier (¿) used in LARDO based mode decision does not consider the correlations between layers. Consequently, ¿ is not optimal for multi-layer scenario in error resilient coding. In this paper, modified ¿ selection method in multi-layer scenario, for mode decision of LA-RDO based error resilient coding is proposed. Simulation results show that significant PSNR gain is reportedly achieved for various tested sequences compared to JSVM 9.8 with quality and spatial scalabilities for IPPP coding, as well as for hierarchical B picture coding.
Keywords :
video coding; IPPP coding; error propagation; error resilient coding; hierarchical B picture coding; loss-aware rate distortion optimized macroblock mode decision; multilayer Lagrange multiplier selection; scalable video coding; transmission errors; Error correction; Lagrangian functions; Multimedia communication; Nonhomogeneous media; Performance loss; Rate distortion theory; Rate-distortion; Scalability; Static VAr compensators; Video coding; Lagrange multiplier; error concealment; error resilient coding; loss-aware rate distortion optimization; scalable video coding;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Advanced Communication Technology (ICACT), 2010 The 12th International Conference on
Conference_Location :
Phoenix Park
ISSN :
1738-9445
Print_ISBN :
978-1-4244-5427-3
Type :
conf
Filename :
5440336
Link To Document :
بازگشت