<![CDATA[Background: Skin diseases caused by fungal and viral infections, such as ringworm, chickenpox, and shingles, often exhibit similar visual patterns in their early stages, making manual classification difficult and potentially leading to misdiagnosis. Objective: This study proposes an image-based skin disease classification approach that combines visual feature extraction with machine learning algorithms.Methods: Feature extraction is performed using pretrained models (Inception-v3, VGG-16, and VGG-19) to generate image embeddings. The extracted features are classified using logistic regression, support vector machine (SVM), and neural network models via the Orange Data Mining platform.Results: Performance evaluation using AUC, classification accuracy (CA), F1 score, precision, recall, and Matthews correlation coefficient (MCC) shows that the combination of Inception-v3 and SVM achieves the best performance. Pretrained feature extraction effectively improves the accuracy of machine learning-based classification. Conclusion: Combining pretrained feature extraction with machine learning provides an efficient and accurate approach to skin disease classification, with strong potential for development into an early-stage clinical decision-support system.]]>
Copyrights © 2026
