<![CDATA[This research implements YOLO v11 for image-based waste detection and classification to improve waste management efficiency. The model recognizes four categories of waste: inorganic, organic, hazardous and residual. The training results showa mAP@0.5 of 0.989 and a maximum F1 of 0.98 at an optimal confidence level of 0.669. The model had high precision on the Organic (0.995) and B3 (0.991) classes, but faced difficulties in classifying the Residue category. The confusion matrix revealed most of the predictions were accurate, despite some misclassification. The model also showed stable performance under various lighting and background conditions. With this reliability, YOLO v11 can be applied in automated sorting systems to improve recycling efficiency and support sustainable environmental management, although further improvements to data augmentation and class weight adjustment are still needed.]]>
Copyrights © 2025
