Fractional differentiation-based variational level set model for noisy image segmentation without contour initialization

Noisy image segmentation is a hot topic in image analysis. In this paper, we present a novel methodology for tackling this issue through the integration of fractional differentiation in the frequency domain with a variational level set model (VLSM), which eliminates user-selected initial contours by incorporating the convex energy function. Additionally, the fractional differentiation reduces noises while preserving more detail information. Experiments on synthetic and real noisy images demonstrate that our proposed model surpasses other denoising VLSMs in terms of noise reduction, segmentation accuracy, and efficiency.