Methods of representing and analysing images

摘要

A method of representing an image comprises deriving at least one descriptor based on colour information and colour interrelation information for at least one region of the image, the descriptor being derived using values of pixels in said region and having at least one descriptor element, wherein at least one descriptor element for a region, optionally each of a plurality of descriptors elements for a region or all the descriptor elements for a region, is or are derived using only a subset of pixels in said region and/or wherein the calculations used to derive the descriptor elements are such that the resulting number of descriptor elements is less than the number of pixels used to derive the descriptor elements and/or wherein at least one descriptor element for a region is derived using a non-wavelet transform.