Title :
Variable-Length Source Compression Using Successive Refinement and Non-linear Graph-Based Codes
Author :
Ramirez-Javega, Francisco ; Lamarca, Meritxell
Author_Institution :
Univ. Politec. de Catalunya, Barcelona, Spain
Abstract :
In this paper, the authors propose a variable length source code that achieves very good performance even for short sequences and has low encoding complexity, although it does not achieve the theoretical limit. Furthermore, the performance can be predicted for infinite codeword length with close form expressions, and has been proven to be good even when a suboptimum decoder with very low complexity is employed at the decoder side.
Keywords :
decoding; graph theory; sequences; source coding; infinite codeword length; low encoding complexity; nonlinear graph based codes; short sequences; suboptimum decoder; successive refinement; variable length source code; variable length source compression; Adaptation model; Bit error rate; Complexity theory; Data compression; Decoding; Encoding; Entropy;
Conference_Titel :
Data Compression Conference (DCC), 2011
Conference_Location :
Snowbird, UT
Print_ISBN :
978-1-61284-279-0
DOI :
10.1109/DCC.2011.77