Method of Medical Image Segmentation based on Chan-Vese Model

During the imaging process, medical images will be affected by other undesirable factors such as the noise of other human tissues or imaging equipment, resulting in images showing low resolution, weak boundaries, severe noise, and uneven light distribution, which increases the difficulty of accurately segmenting the target. The model of Chen Vese is a classical geometric active contour model. It can effectively divide images with weak and intermittent borders, with strong resistance to noise, but it has poor segmentation effect for medical images with irregular gray level. Considering this, this paper utilizes top cap transformation and bottom cap transformation to do preliminary processing of medical images. The new image is a linear combination of top-hat transformation and bottom-hat transformation. Through reasonable selection of parameters, while enhancing the the image’s great difference, the gray level of the image becomes more uniform. Then the model of Chen Vese which is furthered is utilized for dividing the medical image. From the experimental results, the new method improves the curve evolution speed and can segment medical images quickly and effectively.