A Hybrid Skin Lesions Segmentation Approach Based on Image Processing Methods

Presently image segmentation remains the most crucial stage in processing system. The main idea ofimage is to partition or divide a random into several partitions depending on problem solve. In this paper, we will be presenting new method of skin cancer detection based Otsu’s thresholding algorithm and markercontrolled watershed method. This hybridization process first all started by segmenting input using fuzzy c-means which clustering that gives possibility pixel belong one more clusters. After that, apply multi-Otsu separates pixels an variety classes intensity gray levels. next step proposed marker-controlled watershedalgorithm divides homogenous areas regions edge-detection concepts including mathematical morphology. technique was applied experienced images different types were collected gathered from web also Kaggle dataset. To assess worth achieved results, used evaluation metrics like dice coefficient, sensitivity, specificity as well Jaccard similarity have shown good satisfactory results.