Hybrid Technique For Skin Pimples Image Detection and Classification

Abstract

Skin pimples are formed when the follicles of the hair follicles are closed under the skin, or the sebaceous glands are closed, which causes them to swell and appear in several abnormal forms, and a red color contrary to the skin color. In this article, a hybrid Method consists of Naive Bayes Classifier (NBC) and image processing has been developed for pimples detection and classification. An image of the skin area affected with pimples is being considered as the ROI (region of interest). This is taken into the experiment, which results in automatic markings of the pimples and thereby extracting their features and performing the classification of the pimples. The presence of pimples in different parts of the face or body has different indications of skin diseases and might not be dangerous but depends on severity and leaves scar. Detecting different types of pimples lesions is very important in both diagnosis as well as management. To access pimples, clinicians and dermatologists use methods such as ordinary flash photography and direct visual assessment, which is time-consuming. The classification of various pimples types considered in this work includes- pimples Cystic, pimples Excoriated, and pimples Pustular. The experimental outcome can be brief with 93.42 % accuracy by using 40 images of each type.