Downloads & Free Reading Options - Results

Data representation, which is often overlooked in many image processing and analysis applications, is as critical as the algorithms applied to that data. When data is represented properly, simple algorithms can be much more powerful than sophisticated and often complex algorithms applied to an improper representation. In an image, the useful information is generally mixed in with irrelevant information or noise and often it is difficult for the computer to separate the useful information from the large volume of irrelevant data without destroying much of the useful data. To provide a solid foundation for a good solution to this problem, multiresolution decomposition and synthesis approaches have been developed which decompose raw image data into a set of partial information channels, where each channel represents a certain modality (or aspect) of the raw image. The channels can then be processed individually or cooperatively with a wide variety of results possible.