Matching images for radiotherapy verification using genetic algorithms

The paper presents a novel method for aligning images used in the radiotherapy treatment of cancer patients. During the application of such treatment it is important that only the affected areas are exposed to radiation. This area is typically prepared by taking a low energy X-ray image of the patient, followed by the identification of the area to be treated by a radiotherapy consultant. This template image is then used to position lead blocks to protect the surrounding healthy area. After each course of treatment it is important to verify the treatment has been successfully applied. This is achieved by comparing the reference image with the image of the area irradiated. This is currently manually performed and requires a great deal of expertise as well as time. The paper reports an automated image alignment and verification procedure based upon evolutionary computation