mAP
مقياس mAP
Also known as: mean Average Precision
mAP (mean Average Precision) is the headline accuracy metric for object detection. For each class, average precision is the area under the precision-recall curve; mean averages across all classes. mAP@0.5 evaluates at a single IoU threshold; mAP@0.5:0.95 averages across thresholds 0.5 to 0.95 in 0.05 steps and is the harder…
Definition
mAP (mean Average Precision) is the headline accuracy metric for object detection. For each class, average precision is the area under the precision-recall curve; mean averages across all classes. mAP@0.5 evaluates at a single IoU threshold; mAP@0.5:0.95 averages across thresholds 0.5 to 0.95 in 0.05 steps and is the harder COCO standard. FI Tech reports both numbers in customer acceptance tests: PPE classes consistently exceed 0.92 mAP@0.5 on Saudi-construction validation sets, with hard-hat at ~0.99 and reflective-vest at ~0.94 after fine-tuning on local-uniform variants.
