An efficient search of binary tree for huffman decoding based on numeric interpretation of codewords

This paper presents a new method of Huffman decoding which gives a significant improvement of processing efficiency based on the reconstruction of an efficient one-dimensional array data structure incorporating the numeric interpretation of the accrued codewords in the binary tree. In the Proposed search method, the branching address is directly obtained by the arithematic operation with the incoming digit value eliminating the compare instruction needed in the binary tree search. The proposed search method gives 30% of improved Processing efficiency and the memory space of the reconstructed Huffman table is reduced to one third compared to the ordinary ‘compare and jump’ based binary tree. The experimental result with the six MPEG-2 AAC test files also shows about 198% of performance improvement compared to those of the widely used conventional sequential search method.