Research on FCM algorithm in the 3D visualization system of medical images

摘要

FCM is a fuzzy segmentation based on overall situation, is typically applied in data mining and pattern recognition. In this paper, the segmentation of brain CT is achieved through FCM clustering algorithm in three-dimensional medical image visualization system, the organization in brain CT processed with FCM clustering can be well identified. However, the connectivity of brain organization is severely damaged. In view of this situation, it is proposed that the object in the brain image through clustering be judged by classification of its neighbor domain. The result shows that this method brings a significant improvement in the problem of organization connectivity brought by FCM clustering. Judging the brain image through FCM clustering by classification of its neighbor domain, a brain CT image of better organization integrity and connectivity can be got.