VITILIGO DETECTION USING MACHINE LEARNING

Abstract

Skin disorders are widespread worldwide, encompassing various conditions such as skin cancer, vulgaris, ichthyosis, and eczema. Among these, vitiligo stands out as it can appear anywhere on the body, including the oral cavity, and significantly affect overall health, leading to cognitive issues, hypertension, and mental health problems. Traditional diagnostic methods employed by dermatologists, like biopsy, blood tests, and patch testing, have limitations, particularly in cases where lesions progress from macules to patches. To address this, ML and DL models have surfaced as prominent technologies. to expedite diagnosis. This research introduces a Deep Learning-based model specifically designed for anticipating and categorizing vitiligo in healthy skin. Leveraging a pre-trained Inception V3 model, image features are extracted and utilized alongside classifiers such as naive Bayes, convolutional neural network (CNN), random forest, and decision tree. Evaluation metrics including accuracy, recall, precision, and mAP (mean average precision) and F1-score are employed. Our result section confirms with the help of Table.1 that for overall performance and specific results for melanocytes and disease detection, Inception V3 coupled with random forest exhibits exceptional performance across all metrics, achieving a recall of 0.76, precision of 0.81, F1-score of 0.77, mAP@.5:.95 of 0.56, and mAP@.5 of 0.82 for “All” class comprising of 23 targets,  recall of 0.76, precision of 0.82, F1-score of 0.75, mAP@.5:.95 of  0.58, and mAP@.5 of 0.84 for “Melanocytes” class covering 13 targets and recall of 0.78, precision of 0.80, F1-score of 0.79, mAP@.5:.95 of 0.58, and mAP@.5 of 0.83 for “Diseased” class covering 10 targets. Although Inception V3 combined with the decision tree classifier shows slightly lower performance, it still achieves respectable results. Notably, Inception V3 coupled with random forest demonstrates superior performance across most metrics.