20、计算机视觉中的目标检测：从YOLOv3到YOLOv7

计算机视觉中的目标检测：从YOLOv3到YOLOv7

1. 目标检测中的平均精度计算

在目标检测中，平均精度（mAP）是评估模型性能的重要指标。以下是计算平均精度的代码实现：

# a dictionary for true_bboxes number on each image
for key, val in amount_bboxes.items():
    amount_bboxes[key] = torch.zeros(val) #===================line 15 
detections.sort(key=lambda x: x[2], reverse=True) #==========line 16 
TP = torch.zeros((len(detections))) 
FP = torch.zeros((len(detections))) 
total_true_bboxes = len(ground_truths) #total true_bboxes in class c 
if total_true_bboxes == 0:
    continue # If no true_bboxes for class c, skip the following codes 
for detection_idx, detection in enumerate(detections): 
    #===line 22
    trueBboxesOnllmg = [ #Pick out all true bboxes on an image 
        bbox for bbox in ground_