A novel competitive learning technique for the design of variable-rate vector quantizers with reproduction vector training in the wavelet domain

This paper presents a novel competitive learning algorithm for the design o f variable-rate vector quantizers (VQs). The algorithm, termed variable-rat e competitive learning (VRCL) algorithm, designs a VQ having minimum averag e distortion subject to a rate constraint. The VRCL performs the weight vec tor training in the wavelet domain so that required training time is short. In addition, the algorithm enjoys a better rate-distortion performance tha n that of other existing VQ design algorithms and competitive learning algo rithms. The learning algorithm is also more insensitive to the selection of initial codewords as compared with existing design algorithms. Therefore, the VRCL algorithm can be an effective alternative to the existing variable -rate VQ design algorithms for the applications of signal compression.