An Automatic Threshold Segmentation and Mining Optimum Credential Features by Using HSV Model

In this present study a perfect outcome of skin lesion in the computerized image analysis is used to segment the abnormal layers on the skin. The dermatologist finds difficult for easy identification of skin lesion. A computational tool should be developed to assist the dermatologist for diagnosis. This paper reports the differentiation of segmentation with various techniques. The review is made with related works to the current proposed method as a comparative study with plenty of fundamental steps like image acquisition, pre-processing and segmentation. In this work, the asymmetric pattern extractions from the dermoscopic images are segmented by the HSV segmentation to find the contour image. An automatic segregation of RGB–HSV is incorporated in the masked threshold on the proposed system which segments the lesion. The techniques involved in each stage are perfectly explained. From the state of RGB input and handling of pre-processing and segmentation were evaluated effectively. The outcome of this result is compared with other segmentation techniques to improve the result. The proposed performance measures between Ground Truth image and Segmented Image provides best-offered values of accuracy up to 96% for PH² dataset and 95% for ISIC 2016 Dataset.