Current image segmentation methods often suffer from issues such as low accuracy, slow processing speed, and inadequate robustness when dealing with images with inhomogeneous noise and intensity. To resolve these issues, we propose a fast image segmentation algorithm based on a grayscale morphological edge differential fitting model. By utilizing morphological erosion and dilation operations, our model matches the differential image intensity inside and outside the contour. The grayscale morphological operator extracts local image information, which can effectively segment images with intensity inhomogeneity. Since the edge differential fitting function is replaced by the image grayscale morphology, it reduces the need for updates during level set evolution, thereby lowering the CPU runtime and complexity. Experimental results indicate that our model demonstrates fair robustness to noise interference and initial contours. Compared with active contour models (ACMs) and deep learning methods, our model exhibits superior segmentation accuracy while remaining robust to initial contours.
Jian Su (Sat,) studied this question.