Transform coding based on vector quantization using a fuzzy-possibilistic neural network

In this paper, the fuzzy-possibilistic c-means (FPCM) is embedded into a 2-D Hopfield neural network named Fuzzy-Possibilistic Hopfield Net (FPHN) in order to generate optimal solution for Vector Quantization (VQ) in Discrete Cosine Transform (DCT) and Hadamard Transform (HT) domains. The information transformed by DCT or HT was separated into DC and AC coefficients. Then, the AC coefficients were trained using the proposed methods to generate better codebook based on VQ. The energy function of FPHN is defined as the fuzzy membership grades and possibilistic typicality degrees between training samples and codevectors. A near global-minimum codebook in frequency domains can be obtained when the energy function converges to a stable state. Instead of one state in a neuron for the conventional Hopfield nets, each neuron occupies 2 states called membership state and typicality state in the proposed FPHN. The simulated results show that valid and promising codebook can be generated in DCT and Hadamard domains using the FPHN.