Psycho-visually Tuned Area-morphology Tools for Improved Image Compression

In video compression, image pre-processing is used to improve the overall compression performance. A key requirement of any such system is the ability to remove noise and other higher frequency details while preserving the visual quality of the images. This paper presents a new approach to pre-filtering, based on area morphology. Both the standard area and a new constraint, based on the power within the granule, are investigated and psycho-visual experiments used to determine the highest value of each constraint for which the filtering is visually lossless. The compressibility of the pre-processed images, assessed by a number of compression methods, is significantly improved thus demonstrating the advantages of the area morphology methods, with the power constraint providing the greatest improvement over the original images.