The alphabet-constrained theory of data compression provides a specific methodology for obtaining explicit data compression system designs, which is in sharp contrast with rate distortion theory and the usual intuitional design methods. Various aspects of alphabet-constrained data compression were investigated, including preposterior analysis, adaptive code generators, vector quantization and the relationships among prediction, quantization and stationarity. Simulations of several coder designs for speech and image sources were performed.
Reports by Author
