Multi-subspace tree-structured vector quantizer design algorithms

Abstract

To reduce further the searching complexity and memory requirement of tree-structured vector quantization (TSVQ), novel multisubspace TSVQ design algorithms and encoding techniques are proposed. The proposed multisubspace TSVQ design algorithms perform the vector quantization in the spatial domain while using specially designed subspace distortions in the transform domain as cost functions for the optimization process. The dimensionality and basis of subspace distortions are selected based on the local statistics of the partition associated with each nonterminal node of the tree. Experimental results show that extremely low subspace dimension can be used in multisubspace TSVQ based on a fixed transform domain to obtain a performance similar to that of the conventional TSVQ or single subspace TSVQ. The proposed generalized multisubspace TSVQ design algorithm is a general TSVQ design algorithm and can be utilized as a fast codebook design algorithm.< >