Table 10 Experimental results of the SOTA model on the LOL dataset.
From: Complex dark environment-oriented object detection method based on YOLO-AS
Method | Params(M) | Flops (G) | map@50 (%) | map@50:95 (%) | Speed (ms) |
|---|---|---|---|---|---|
YOLOv3 | 61.5 | 156 | 64.96 | 32.73 | 21.80 |
YOLOv4 | 63 | 180 | 65.10 | 33.10 | 24.30 |
YOLOv5s | 7.2 | 16.5 | 60.37 | 33.54 | 8.20 |
YOLOv7 | 37 | 105 | 61.42 | 33.76 | 12.30 |
YOLOv8n | 3.2 | 8.2 | 61.03 | 33.59 | 5.70 |
YOLOv11 | 48 | 135 | 68.21 | 35.72 | 13.20 |
DETR | 41 | 86 | 66.79 | 35.17 | 38.10 |
Swin Transformer | 48 | 145 | 68.32 | 36.43 | 41.80 |
Faster R-CNN | 137 | 207 | 65.48 | 34.41 | 53.30 |
RetinaNet | 36.7 | 97 | 65.37 | 34.59 | 52.70 |
EfficientDet | 3.9 | 2.5 | 62.12 | 34.16 | 9.10 |
YOLO-AS | 43.0 | 105.6 | 67.81 | 35.88 | 15.40 |
