Improved Model for Skin Illnesses Classification Utilizing Gray-Level Co-occurrence Matrix and Convolution Neural Network

Abstract

Skin conditions are more common than other illnesses. Skin issues can be brought on by viruses, bacteria, allergies, fungi, etc. The detection of skin diseases has been made better by lasers and photonics since it is now quicker and more precise. However, such a diagnosis is pricey. A system for automated dermatology screening is built with the use of computer vision. Using the Gray Level Co-occurrence Matrix (GLCM) and Convolution Neural Network (CNN) provides an improved model for accurately diagnosing skin disorders. In order to classify skin photos, CNN is used by the model to extract features from the images using GLCM. The high-level features utilizing the statistical features were retrieved via GLCM separately because the photos utilized in the research are for skin conditions. Once merged, these features created a high-accuracy categorization. Two distinct classification processes are used to categorize photos into 13 diseases: First, the Deep Neural Network (DNN) classifier obtains 96.69% accuracy, 96.2% recall, 96.2% precision, and 96.2% F1-score in terms of performance evaluation measures. Second, accuracy, recall, precision, and F1-score are the performance evaluation metrics for the Multiple Support Vector Machine (MSVM) classifiers. The model outperforms other cutting-edge models in terms of accuracy and effectiveness when compared to them. This work thus indicates the capability of GLCM and CNN for the classification of skin diseases and their prospective uses in the healthcare sector.