A Research On Segmentation Method For Medical Image Using Graph Cut Technology

As the key technology of computer aided diagnosis system, medical image segmentation problem has always been a hot topic in the field of image segmentation problem. Owing the diversity and complexity of medical images, the traditional segmentation method cannot be good for medical image segmentation, sometimes even make fatal error. Especially the segmentation of soft tissue organs is more difficult. Therefore, according to the characteristics of abdominal CT image sequences, at first, this paper used a hybrid method which based on three-dimensional region growing combined with image morphology to extract liver contour, and adopt the automatic method of threshold detection setting growth standards; than image segmentation algorithm based on graph cut was applied to the extraction of liver contour, through the experimental comparison to choose the suitable weighting function to get the optimal segmentation result. Experiment shows that the result of the method based on graph cut segmentation is slightly better than three-dimensional region growing. The result has a high accuracy and can be used as a data set of three-dimensional reconstruction.