操纵 YOLOv8 结果

Question

如何操作空结果？ （即没有检测到）

例如：

608x800 (no detections), 320.5ms

代码：

    for box in boxes:
        if box.cls == 0:
            print("large bottled water")
        elif box.cls == 1:
            print("medium bottled water")
        elif box.cls == 2:
            print("medium soda")
        elif box.cls == 3:
            print("medium thick bottle")
        elif box.cls == 4:
            print("small bottled water")
        elif box.cls == 5:
            print("small soda")
        elif box.cls == 6:
            print("small thick bottle")
        elif box.cls == 7:
            print("small thick bottle")
        elif box.cls == 8:
            print("xs soda")
        elif box.cls == 9:
            print("xs thick bottle")

我只尝试过这个：

if box.cls == "":
   print("no detection")

我阅读了文档但找不到解决方案

Answer 1

在这种情况下，我会在迭代之前检查

result

是否为空（未检测）。从 torch.Tensor 获取普通

cls

值：

cls = int(box.cls.item())

from ultralytics import YOLO

model = YOLO('yolov8s.pt')
results = model(source='image.jpg') # or different type of source

for result in results:
  if result:
    for box in result.boxes:
      cls = int(box.cls.item())
      if cls == 15:
          print("large bottled water")
      elif cls == 1:
          print("medium bottled water")
      elif cls == 2:
          print("medium soda")
  else:
    print("no detection")