Modified stack search algorithm for VQ-based tree encoding

A modified stack algorithm for VQ-based tree encoding on image compression applications is presented. The encoding process in vector quantization (VQ) is computationally intensive due to its massive search computation operations. There are, generally, two approaches developed to overcome this problem. One is a fast full search approach and the other is a tree search approach (or tree-structured VQ). A modified stack algorithm which is based on a perfectly regular data structure, “stack”, for binary tree search is proposed. Some appropriate metric functions for stack tree search in the VQ-based codebook tree are also suggested. The experimental results that show the improvement over the binary search are reported