Object Detection on AODRaw 1.0 (test)
34.8APCascade RCNN
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Cascade RCNNBackbone=ConvNeXt-T, Pre-Train=RAW, Fine-Tune=RAW2024.11 | 34.8 | 53.3 | 36.7 | 20.6 | 42.8 | 52.5 | 37.7 | 32.1 | 36.1 | 28.4 | |
| Cascade RCNNBackbone=ConvNeXt-T, Pre-Train=sRGB, Fine-Tune=sRGB2024.11 | 34 | 52.7 | 36.3 | 19.3 | 40.8 | 52.1 | 37 | 31.5 | 32.9 | 27.2 | |
| Cascade RCNNBackbone=ConvNeXt-T, Pre-Train=sRGB, Fine-Tune=RAW2024.11 | 33.7 | 52 | 35.9 | 18.6 | 41.7 | 51.3 | 36.8 | 31.3 | 31.3 | 27.2 | |
| Cascade RCNNBackbone=Swin-T, Pre-Train=RAW, Fine-Tune=RAW2024.11 | 32.2 | 50.5 | 33.8 | 17.9 | 40.5 | 49.7 | 35.5 | 30 | 29.5 | 25.1 | |
| GFocalBackbone=ConvNeXt-T, Pre-Train=sRGB, Fine-Tune=sRGB2024.11 | 32.1 | 49.9 | 33.6 | 18.7 | 39.9 | 49.5 | 35.2 | 30.3 | 31.8 | 26 | |
| GFocalBackbone=ConvNeXt-T, Pre-Train=RAW, Fine-Tune=RAW2024.11 | 32.1 | 50.4 | 33.4 | 17.7 | 40.6 | 49.6 | 35.8 | 29.9 | 32.8 | 27.1 | |
| Cascade RCNNBackbone=Swin-T, Pre-Train=sRGB, Fine-Tune=sRGB2024.11 | 32 | 50.2 | 34 | 17.5 | 40.1 | 46.3 | 35.4 | 30 | 28.2 | 25 | |
| Cascade RCNNBackbone=Swin-T, Pre-Train=sRGB, Fine-Tune=RAW2024.11 | 31.7 | 49.8 | 32.8 | 17.7 | 39.7 | 47.8 | 35.3 | 29.8 | 28.6 | 23.9 | |
| GFocalBackbone=ConvNeXt-T, Pre-Train=sRGB, Fine-Tune=RAW2024.11 | 31.5 | 50 | 32.9 | 17.9 | 39.5 | 48.4 | 34.9 | 29.4 | 32.2 | 26.7 | |
| GFocalBackbone=Swin-T, Pre-Train=RAW, Fine-Tune=RAW2024.11 | 30.7 | 49.7 | 31.8 | 17.2 | 39.4 | 47.4 | 33.7 | 28.6 | 28.5 | 25.3 | |
| Faster RCNNBackbone=ConvNeXt-T, Pre-Train=RAW, Fine-Tune=RAW2024.11 | 30.2 | 52.3 | 31 | 17 | 39.1 | 46.9 | 33.8 | 27.7 | 30.2 | 26.6 | |
| GFocalBackbone=Swin-T, Pre-Train=sRGB, Fine-Tune=sRGB2024.11 | 30.1 | 48.9 | 30.6 | 16.3 | 38 | 44.4 | 32.7 | 28.1 | 28.2 | 24.5 | |
| GFocalBackbone=Swin-T, Pre-Train=sRGB, Fine-Tune=RAW2024.11 | 29.9 | 48.2 | 30.6 | 16.3 | 38.3 | 45 | 33.1 | 27.6 | 29 | 23.8 | |
| Faster RCNNBackbone=ConvNeXt-T, Pre-Train=sRGB, Fine-Tune=sRGB2024.11 | 29.7 | 51.7 | 30.1 | 17.1 | 37.3 | 45.4 | 33.1 | 28.3 | 27.1 | 24.4 | |
| Faster RCNNBackbone=ConvNeXt-T, Pre-Train=sRGB, Fine-Tune=RAW2024.11 | 29.4 | 51.3 | 29.6 | 16.3 | 37.6 | 44.4 | 32.7 | 27.3 | 29.2 | 24.6 | |
| Faster RCNNBackbone=Swin-T, Pre-Train=RAW, Fine-Tune=RAW2024.11 | 28.6 | 50.2 | 28.5 | 15.6 | 36.9 | 43.1 | 32.1 | 26.7 | 27.6 | 23.2 | |
| Faster RCNNBackbone=Swin-T, Pre-Train=sRGB, Fine-Tune=sRGB2024.11 | 28.4 | 50.1 | 28.8 | 15.6 | 35.9 | 42.6 | 32 | 26 | 27.2 | 23.3 | |
| Faster RCNNBackbone=Swin-T, Pre-Train=sRGB, Fine-Tune=RAW2024.11 | 28.1 | 50 | 28.2 | 16 | 35.7 | 42.6 | 30.7 | 26.5 | 26.2 | 22 | |
| Cascade RCNNBackbone=ResNet-50, Pre-Train=sRGB, Fine-Tune=sRGB2024.11 | 25.6 | 41.4 | 26.4 | 13.7 | 32.4 | 38.3 | 27.3 | 23.8 | 24.7 | 20.4 | |
| GFocalBackbone=ResNet-50, Pre-Train=sRGB, Fine-Tune=sRGB2024.11 | 24.2 | 40.3 | 24.7 | 13.3 | 31.9 | 37 | 26.5 | 22.3 | 24.1 | 21.1 | |
| Faster RCNNBackbone=ResNet-50, Pre-Train=sRGB, Fine-Tune=sRGB2024.11 | 23.3 | 41.3 | 23.7 | 13.1 | 30.8 | 36.4 | 26 | 22 | 24.4 | 19.6 | |
| RetinanetBackbone=ResNet-50, Pre-Train=sRGB, Fine-Tune=sRGB2024.11 | 19.1 | 33.6 | 19.2 | 10.1 | 26.6 | 29.5 | 21.5 | 17.8 | 19.2 | 16.5 | |
| Deformable DETRBackbone=ResNet-50, Pre-Train=sRGB, Fine-Tune=sRGB2024.11 | 16.6 | 31.9 | 15.6 | 7.7 | 23.9 | 30.1 | 18.3 | 15.2 | 16.4 | 13.1 | |
| Sparse RCNNBackbone=ResNet-50, Pre-Train=sRGB, Fine-Tune=sRGB2024.11 | 15.6 | 28.3 | 15 | 7.2 | 22.1 | 28.9 | 17.9 | 15 | 14.6 | 12.6 |