Identification of skin disease using Gray Level Co-occurrence Matrix and Support Vector Machine / Muhammad Amsyar Omar

Chronic skin disease such as psoriasis and eczema are two of the most common skin disease affecting the human body especially in Malaysia. These skin diseases can cause serious health and financial effects if not recognised and treated early. Early detection of disease severity, as well as advice on...

Full description

Saved in:
Bibliographic Details
Main Author: Omar, Muhammad Amsyar
Format: Thesis
Language:English
Published: 2022
Subjects:
Online Access:https://ir.uitm.edu.my/id/eprint/59384/1/59384.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Chronic skin disease such as psoriasis and eczema are two of the most common skin disease affecting the human body especially in Malaysia. These skin diseases can cause serious health and financial effects if not recognised and treated early. Early detection of disease severity, as well as advice on skincare and medicine, can help keep the condition from worsening. However, the current diagnosis might be time-consuming and expensive. Hence, this project aimed to develop automated skin disease detection focusing on psoriasis and eczema as the common skin disease in Malaysia. To accomplish this, skin disease images were pre-processed to filter and segment the image by enhancing, removing noise, and segmenting the image. Then, the method Gray Level Co-Occurrence Matrix (GLCM) was used to extract features of the skin disease images that could be obtained correctly. The identification of the skin disease is performed in the enhanced images using Support Vector Machine (SVM) classifier. A set of 20 different skin disease images were analysed and utilized, giving an overall accuracy of 90% for skin disease identification. These findings indicate that the proposed system can assist patients and dermatologist in determining the type of disease from an image of the affected region during the early stages of skin disease.