A progressive iterative deformation snake model for segmentation of images with multi-level complexity

We propose in this work a new and efficient active contour technique for segmenting diverse types of images. Our approach presents a new snake deformation algorithm named Progressive Iterative Deformation (PID) method, which iteratively fits control points to construct a series of fitting curves until the object boundary is detected, without the need to solve energy minimization problems. In addition, to accommodate various image types, we propose a new weighted external energy expression that controls segmentation quality based on texture and feature extraction. This method operates over three image complexity classes, adapting to segmentation challenges and image quality. We evaluate the robustness and efficiency of the proposed PID-Snake model across three different dataset classes. Our results demonstrate that the proposed technique improves deformation complexity cost and segmentation performance. In particular, our model is effective and powerful against the difficulties presented by images such as those in the medical field.