Methods to Accelerate a Competitive Learning Algorithm Applied to VQ Codebook Desing

Abstract

Abstract. Codebook design plays a crucial role in the performance of signal processing systems based on vector quantization (VQ). This paper is concerned with methods for reducing the processing time spent by a competitive learning (CL) algorithm applied to VQ codebook design. Using analytical expressions for the number of operations (multiplications, additions, subtractions and comparisons) performed by the CL algorithm, it is shown that almost all the operations are due to the nearest neighbor search (NNS). Simulation results regarding image VQ show that simple modifications introduced in CL lead to considerable number clock cycles savings.