Table 8 Experimental results of different attention mechanisms 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) |
|---|---|---|---|---|---|
NONE | 36.5 | 104.7 | 61.42 | 33.76 | 12.30 |
SE | 39.9 | 104.7 | 60.16 | 33.12 | 13.60 |
ECA | 39.5 | 104.7 | 60.36 | 32.87 | 13.10 |
CBAM | 39.6 | 104.8 | 62.41 | 33.91 | 14.20 |
CA | 39.8 | 104.7 | 60.72 | 32.69 | 14.50 |
MLCA | 39.5 | 104.8 | 62.55 | 34.45 | 13.70 |
ECA_ASPP + SK | 42.8 | 104.8 | 67.81 | 35.88 | 15.40 |
