[Method for segmenting medical image based on anatomical model].

OBJECTIVE: To improve the automatization and reliability of medical image segmentation. METHOD: An anatomical model was built and used to guide the low-level segmentation process. The system architecture was made up of an anatomical model, image processing routines and an inference engine, the interaction of which are governed by a blackboard. RESULT: The result of application of the segmentation for chest CT image was satisfactory and needs less operator intensive. CONCLUSION: This method improves automatization and reliability of the medical image segmentation. Because of the good expansibility, it may serve as a template for knowledge-based processing of medical image.