Detection and remedy of skin diseases using image texture & color features App

Abstract

Skin-related diseases are among the most common and rapidly growing health problems in the world. Because of how complicated the texture, tone, existence of hair, and other distinguishing characteristics are, human skin disorders are the most unexpected and difficult things to automatically identify and analyse. This method will inform us if the condition is acne or skin cancer, and after more processing, it will forecast the severity of the condition that we have designated. A smartphone application called "Detection and Remedy of Skin Diseases using Image Texture and Colour Features App" has a variety of functions to assist users with skin issues. This program aims to provide assistance to patients in remote, economically disadvantaged, and developing areas by allowing them to conduct routine skin inspections using their smartphones. This is made possible by the prevalence of smartphones in today's society and the desire for convenience in performing important tasks. With this program, patients can easily scan and analyze their skin from anywhere, making it a valuable tool for those who may not have access to traditional medical resources. The technology successfully detects a variety of skin issues. This process' three main components are image processing, training, and app creation. During the picture processing stage, we must employ extraction and pre-processing techniques. Mobilenetv2 and Tensorflow custom classifier were two machine learning algorithms used in the classification of skin diseases. Following extensive testing and assessment, these algorithms were selected as the final option. The final machine learning model was then included into a mobile application, giving consumers a quick and simple tool for spotting skin diseases. The creation of sophisticated tools like the one we made is a result of the increased usage of ML techniques in image processing apps in recent years. The process of locating and diagnosing skin illnesses relies heavily on image processing methods such feature extraction, segmentation, and picture enhancement. To diagnose the disease, these procedures are used to separate the afflicted part of the body and examine the skin's colour, texture, and form. In this project, we present an overview of several image processing techniques that may be applied to the detection and diagnosis of skin diseases. This thesis also offers a thorough analysis of numerous skin problems and the methods utilised to spot them. Our goal is to increase the precision and efficacy of diagnosing and treating skin diseases by utilising cutting-edge image processing techniques.