System for automatic realignment of the head in magnetic resonance imaging

The condition of patients with various diseases of the brain, may be monitored over a period of time using magnetic resonance imaging (MRI) techniques. To make an accurate assessment of changes in their condition it is necessary that the patients be repositioned to a very high degree of accuracy. Current techniques rely on the skill of the radiographer to return the head to the same position during each scanning. Various mechanical fixation methods, such as the use of head moulds and indelible markers, have been proposed. The authors describe an approach to automatic patient realignment using a combination of image analysis techniques and symbolic reasoning. A set of key anatomical features are extracted from the images and these are compared using a knowledge-based expert system to measure and then correct the positional inaccuracies. The system is capable of extension for crossmatching images from other modalities