A Novel Approach for Segmentation and Classification of Skin Diseases based on Active Contour Model and Multiclass SVM Classifier

Skin disorders are a major worldwide health concern that requires quick intervention and effective diagnostic methods. For efficient medical assessment and therapy, skin diseases must be correctly identified and classified. This work presents a novel approach for the accurate segmentation as well as categorization of different skin conditions, using the Active Contour Model (ACM), Region Growing (RG), and a Multiclass Support Vector Machine (SVM) Classifier. The suggested method begins with a strong region-growing (RG) technique that makes it easier to find starting regions of interest (ROI) in skin imagery. Then, to improve segmentation preciseness, the ACM is used to fine-tune and define the borders of these regions. A MSVM Classifier is then employed with the features that were retrieved from the partitioned regions as input, enabling a thorough and precise categorization of various skin conditions. To validate the suggested strategy, comprehensive experiments were carried out on a broad dataset that included a wide spectrum of skin conditions. The effectiveness of proposed approach is evidenced by the results, which show excellent performance in both the categorization and segmentation assessments.