Image encoding method and image encoder

摘要

An image signal is transformed by at least two different mathematical transformations to obtain at least two different transformed signals. A selected one of the transformed signals is quantized and coded, the selection being made so as to minimize the size of the coded data. The selection is made by comparing the energy convergence rates of the transformed signals, or by quantizing all of the transformed signals and comparing the amounts of non-zero quantized data, or by quantizing and coding all of the transformed signals and comparing the coded data sizes directly. <IMAGE>