A Novel Image Segmentation Algorithm Based on Hesitant Neutrosophic and Level Set

The first step of dermoscopy image analysis technology is image segmentation, the segmentation result will directly affect the subsequent processing process, aiming at the background noise, blurred edge and gray uneven dermoscopy image, a new image segmentation method combining the hesitant neutrosophic set with level set is proposed. Firstly, the image is transformed into the hesitant neutrosophic image by using the hesitant neutrosophic set theory. The hesitant neutrosophic set image is composed of three subsets (T, I, F), and the hesitant neutrosophic image is used to highlight the target information and edge information of the image. Then, we improved the traditional DRLSE level set, constructed a new edge stop function, and added a gray energy term. Finally, we used the improved DRLSE level set to test the ISIC(2018) dermoscopy image. The Jaccard Index values of the experimental results are all greater than 95%, and the mean square error (MSE), peak signal-to-noise ratio (PNSR) and structural similarity Index (SSIM) all perform well, indicating that the proposed method can accurately and effectively segment the skin mirror images with fuzzy edges and uneven gray scales. It's a foundation for the processing and diagnosis of the subsequent dermoscopy image.