| 摘要 |
An adaptive vector quantization process and quantizer (VQ) using a clustering technique known as AFLC (adaptive fuzzy leader clustering) is disclosed. The quantizer, AFLC-VQ, has been designed to vector quantize wavelet decomposed sub images with optimal bit allocation. The high-resolution sub images at each level have been statistically analyzed to conform to generalized Gaussian probability distributions by selecting the optimal number of filter taps. The adaptive characteristics of AFLC-VQ stem from using self-organizing neural networks with fuzzy membership values of the input samples for upgrading the cluster centroids based on well known optimization criteria. The centroid of each pattern group can represent all the other patterns in that cluster. The entire data set can be indexed into look up tables sent to the user for decoding and viewing the image. By generating look up tables or codebooks and entropy coding of the quantizer output, AFLC-VQ can compress images at 100:1 or greater for transmitting while reconstructing the images with high fidelity.
|
