Image compression method and apparatus

Encoding or compressing video or a sequence of images where a desired target number of bits is determined for each of a plurality of image areas in an image to be encoded. The target bitrate is determined from a weight associated to the current image area. Determining this weight comprises: normalising an actual number of bits generated from encoding a corresponding area in a previously encoded image to produce a normalised number of bits with respect to the previously encoded image; determining a normalised weight 1/4 from the normalised number of bits for each of a set of previously encoded images; combining the normalised weight 1/4 with the weight 3/4 used to encode the corresponding area of the previous image. The problem solved is the setting of different target bit rates for different areas of the image, such as tiles or groups of tiles. The target bitrate is set based on a weight associated to a tile or tile group and is based on a normalised weight calculated in part from the actual bitrate of collocated tiles in previous images and on the previous weight used for the collocated tiled in the previous image.