<![CDATA[Monkeypox is an infectious disease characterized by skin lesions that are often difficult to distinguish from other pox-related conditions, which complicates diagnosis in resource-limited settings. This study aims to implement YOLOv7 for detecting monkeypox lesions in dermatological images and to evaluate its accuracy. The dataset consisted of 1,500 annotated images resized to 512×512 pixels, monkeypox was used as the target class, while chickenpox and cowpox were included as comparison/non-target classes to support the differentiation of lesions during model training and evaluation. The YOLOv7 model was trained for 50 epochs using default configurations and a transfer learning approach, with a data split of 70% for training, 20% for validation, and 10% for testing. Training results showed an mAP@0.5 of 89.1% and an mAP@0.5:0.95 of 59.2%. Meanwhile, on the testing stage using original (non-augmented) data, the model performance decreased, achieving an mAP@0.5 of 75.3% and an mAP@0.5:0.95 of 44.9%.]]>
Copyrights © 2025
