Level set method for image segmentation model based on improved signed pressure force function
Jiaxin Wang, Jing Liu, Teng Wu · 2021
A new level set model is proposed to solve the problems of selective binary and gaussian filtering regularized level set (SBGFRLS), which cannot segment the image with weak boundary or uneven gray scale. A new signed pressure force function (SPF) which fuses global and local statistical information is constructed by introducing the generalized global mean gray value and adding local image gray value. In order to speed up the curve evolution, the velocity function is improved. The experimental results show that the new model can effectively segment targets with fuzzy boundaries and multi-target images with different gray scales, and has some anti-noise properties.