DocumentCode :
610064
Title :
Compressing Huffman Models on Large Alphabets
Author :
Navarro, G. ; Ordonez, A.
Author_Institution :
Dept. of Comput. Sci., Univ. of Chile, Santiago de Chile, Chile
fYear :
2013
fDate :
20-22 March 2013
Firstpage :
381
Lastpage :
390
Abstract :
A naive storage of a Huffman model on a text of length n over an alphabet of size σ requires O(σlog n) bits. This can be reduced to σ logσ + O(σ) bits using canonical codes. This overhead over the entropy can be significant when σ is comparable to n, and it also dictates the amount of main memory required to compress or decompress. We design an encoding scheme that requires σlog log n+O(σ+log2 n) bits in the worst case, and typically less, while supporting encoding and decoding of symbols in O(log log n) time. We show that our technique reduces the storage size of the model of state-of-the-art techniques to around 15% in various real-life sequences over large alphabets, while still offering reasonable compression/decompression times.
Keywords :
Huffman codes; computational complexity; decoding; entropy; formal languages; Huffman model compression; canonical codes; compression times; decoding; decompression times; encoding scheme; entropy; large alphabets; Computational modeling; Data structures; Decoding; Encoding; Entropy; Vectors; Vegetation; Huffman compression; compressed data structures;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Data Compression Conference (DCC), 2013
Conference_Location :
Snowbird, UT
ISSN :
1068-0314
Print_ISBN :
978-1-4673-6037-1
Type :
conf
DOI :
10.1109/DCC.2013.46
Filename :
6543074
Link To Document :
بازگشت