| Summary: | 碩士 === 國立台灣工業技術學院 === 管理技術研究所 === 85 === This paper first transforms the Huffman treeinto a single--side
growing Huffman tree, then presents a memory--efficientdata
structure to represent the single--side growing Huffman tree,
whichrequires $O((n+d)\lceil \log_{2} n \rceil)$--bits memory
space, where $n$ isthenumber of source symbols and $d$ is the
depth of the Huffman tree.Based on the proposed data structure,
we present an$O(d)$--timeHuffman decoding algorithm.Using the
sameexample, thememory required in our decoding algorithm is
much less than that ofHashemian in 1995. We finally modify our
proposed data structure slightly todesign an$O(1)$--time
parallel Huffman decoding algorithmon a concurrent read
exclusive write parallel random--access machine(CREW PRAM) using
$d$ processors.
|
|---|
