Oriented Object Detection on DOTA v1.0 (test)
83.52SVPPYOLOE-R-m
Evaluation results
| Method | Links | ||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| PPYOLOE-R-mPretrain=IN, Backbone=CSPRepResNet-m, MS=true2022.12 | 83.52 | 79.71 | 88.63 | 84.45 | 56.27 | 79.12 | 86.16 | 88.77 | 90.81 | 88.01 | 88.39 | 70.41 | 61.44 | 77.65 | 77.7 | 74.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PPYOLOE-R-xPretrain=IN, Backbone=CSPRepResNet-x, MS=true2022.12 | 83.34 | 80.73 | 88.45 | 84.46 | 60.57 | 77.7 | 85.36 | 88.97 | 90.78 | 88.53 | 87.47 | 69.26 | 65.96 | 77.86 | 81.36 | 80.93 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PPYOLOE-R-lPretrain=IN, Backbone=CSPRepResNet-l, MS=true2022.12 | 83.13 | 80.02 | 88.4 | 84.75 | 58.91 | 76.35 | 86.1 | 88.79 | 90.87 | 88.74 | 87.71 | 67.71 | 68.44 | 77.92 | 76.17 | 76.35 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PPYOLOE-R-sPretrain=IN, Backbone=CSPRepResNet-s, MS=true2022.12 | 82.57 | 79.42 | 88.93 | 83.95 | 56.6 | 79.4 | 85.89 | 88.64 | 90.87 | 87.82 | 87.54 | 68.94 | 63.46 | 76.66 | 79.19 | 70.87 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PPYOLOE-R-xPretrain=IN, Backbone=CSPRepResNet-x, MS=false2022.12 | 82.4 | 78.28 | 89.49 | 79.7 | 55.04 | 75.59 | 85.2 | 88.35 | 90.76 | 85.69 | 87.7 | 63.17 | 69.52 | 77.09 | 75.08 | 69.38 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LSKNet-SArchitecture=Two-stage, Pre-trained=ImageNet, Parameters=31.0M, FLOPs=161G, EMA finetune=true2024.03 | 82.2 | 81.64 | 89.57 | 86.34 | 63.13 | 83.67 | 86.1 | 88.66 | 90.89 | 88.41 | 87.42 | 71.72 | 69.58 | 78.88 | 81.77 | 76.52 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Oriented R-CNNBackbone=LSKNet-S, Multiscale training and testing=true2024.07 | 82.2 | 81.64 | 89.57 | 86.34 | 63.13 | 83.67 | 86.1 | 88.66 | 90.89 | 88.41 | 87.42 | 71.72 | 69.58 | 78.88 | 81.77 | 76.52 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Oriented R-CNNBackbone=LSKNet-S [12], multiscale=true2024.07 | 82.2 | 81.64 | 89.57 | 86.34 | 63.13 | 83.67 | 86.1 | 88.66 | 90.89 | 88.41 | 87.42 | 71.72 | 69.58 | 78.88 | 81.77 | 76.52 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-R-tinyPretrain=IN, Backbone=RTMDet-tiny backbone, MS=false2022.12 | 82.05 | 75.36 | 89.21 | 80.03 | 47.88 | 69.73 | 83.33 | 88.63 | 90.91 | 86.31 | 86.85 | 59.94 | 62.3 | 74.23 | 71.97 | 57.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DAFNEArchitecture=One-stage, Pre-trained=ImageNet2024.03 | 82.04 | 76.95 | 89.4 | 86.27 | 53.7 | 60.51 | 81.17 | 88.66 | 90.37 | 83.81 | 87.27 | 53.93 | 69.38 | 75.61 | 81.26 | 70.86 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DAFNePretrain=IN, Backbone=ResNet-101, MS=false2022.12 | 82.04 | 76.95 | 89.4 | 86.27 | 53.7 | 60.51 | 81.17 | 88.66 | 90.37 | 83.81 | 87.27 | 53.93 | 69.38 | 75.61 | 81.26 | 70.86 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PPYOLOE-R-mPretrain=IN, Backbone=CSPRepResNet-m, MS=false2022.12 | 81.86 | 77.64 | 89.23 | 79.92 | 51.14 | 72.94 | 84.56 | 88.68 | 90.85 | 86.85 | 87.48 | 59.16 | 68.34 | 73.78 | 81.72 | 68.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PPYOLOE-R-lPretrain=IN, Backbone=CSPRepResNet-l, MS=false2022.12 | 81.85 | 78.14 | 89.18 | 81 | 54.01 | 70.22 | 85.16 | 88.81 | 90.81 | 86.99 | 88.01 | 62.87 | 67.87 | 76.56 | 79.13 | 69.65 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Oriented RCNNPretrain=MAE, Backbone=RVSA-ViTAE-B, MS=false2022.12 | 81.73 | 81.18 | 89.4 | 83.94 | 59.76 | 82.1 | 85.32 | 88.88 | 90.86 | 85.69 | 87.65 | 63.7 | 69.94 | 84.72 | 84.16 | 79.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-R-sPretrain=IN, Backbone=RTMDet-s backbone, MS=false2022.12 | 81.55 | 76.93 | 89.18 | 80.45 | 52.09 | 71.35 | 84.05 | 88.79 | 90.89 | 87.83 | 86.98 | 59.58 | 62.28 | 75.9 | 81.96 | 61.04 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RVSA-OModel=ViT-B, Pre-training=M2023.08 | 81.55 | 80.88 | 89.32 | 84.52 | 62.53 | 81.86 | 86.53 | 89 | 90.76 | 87.67 | 88.29 | 67.31 | 71.75 | 79.63 | 80.25 | 72.25 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LSKNet-TArchitecture=Two-stage, Pre-trained=ImageNet, Parameters=21.0M, FLOPs=124G, EMA finetune=true2024.03 | 81.54 | 81.37 | 89.14 | 84.9 | 61.78 | 83.5 | 85.87 | 88.64 | 90.89 | 88.02 | 87.31 | 71.55 | 70.74 | 78.66 | 79.81 | 78.16 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CSL(EAutoDet-m)Backbone=EAutoDet-m, Two-stage detection framework=false, Multi-scale training=false2022.03 | 81.51 | — | 89.04 | 86.61 | 53.83 | 63.05 | 85.12 | 88.46 | 90.77 | 88.21 | 87.93 | 56.04 | 60.96 | 75.96 | 82.13 | 66.18 | — | — | — | — | 77.05 | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-R-mPretrain=IN, Backbone=RTMDet-m backbone, MS=false2022.12 | 81.48 | 78.24 | 89.17 | 84.65 | 53.92 | 74.67 | 83.99 | 88.71 | 90.85 | 87.43 | 87.2 | 59.39 | 66.68 | 77.71 | 82.4 | 65.28 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-R-mBackbone=CSPNeXt-m, Scale configuration=single-scale training and testing2023.05 | 81.48 | — | 89.17 | 84.65 | 53.92 | 74.67 | 83.99 | 88.71 | 90.85 | 87.43 | 87.2 | 59.39 | 66.68 | 77.71 | 82.4 | 65.28 | — | — | — | — | 78.24 | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-R-tinyPretrain=IN, Backbone=RTMDet-tiny backbone, MS=true2022.12 | 81.47 | 79.82 | 87.89 | 85.7 | 55.83 | 81.28 | 85.12 | 88.91 | 90.88 | 88.15 | 87.96 | 67.29 | 68.59 | 77.71 | 80.5 | 69.96 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CSL(YOLOv5-s)Backbone=YOLOv5-s, Two-stage detection framework=false, Multi-scale training=false2022.03 | 81.38 | — | 88.93 | 76.07 | 50.97 | 59.61 | 84 | 88.19 | 90.81 | 80.52 | 87.63 | 47.38 | 62.85 | 68.27 | 80.63 | 62.74 | — | — | — | — | 74 | — | — | — | — | — | — | — | — | — | — | — | |
| LSKNet-S*Architecture=Two-stage, Pre-trained=ImageNet, Parameters=31.0M, FLOPs=161G, EMA finetune=true2024.03 | 81.37 | 81.85 | 89.69 | 85.7 | 61.47 | 83.23 | 86.05 | 88.64 | 90.88 | 88.49 | 87.4 | 71.67 | 71.35 | 79.19 | 81.77 | 80.86 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LSKNetModel=LSKNet-S, Pre-training=IN2023.08 | 81.37 | 81.85 | 89.69 | 85.7 | 61.47 | 83.23 | 86.05 | 88.64 | 90.88 | 88.49 | 87.4 | 71.67 | 71.35 | 79.19 | 81.77 | 80.86 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-RArchitecture=One-stage, Pre-trained=COCO, Parameters=52.3M, FLOPs=205G2024.03 | 81.3 | 81.33 | 88.01 | 86.17 | 58.54 | 82.44 | 84.82 | 88.71 | 90.89 | 88.77 | 87.37 | 71.96 | 71.18 | 81.23 | 81.4 | 77.13 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CSL(EAutoDet-s)Backbone=EAutoDet-s, Two-stage detection framework=false, Multi-scale training=false2022.03 | 81.3 | — | 88.64 | 84.78 | 51.02 | 62.04 | 84.02 | 88.23 | 90.82 | 85.79 | 87.36 | 52.2 | 65.68 | 73.72 | 74.33 | 65.93 | — | — | — | — | 75.72 | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-R-lPretrain=COCO, Backbone=RTMDet-l backbone, MS=true2022.12 | 81.3 | 81.33 | 88.01 | 86.17 | 58.54 | 82.44 | 84.82 | 88.71 | 90.89 | 88.77 | 87.37 | 71.96 | 71.18 | 81.23 | 81.4 | 77.13 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-RModel=RTM-L, Pre-training=CO2023.08 | 81.3 | 81.33 | 88.01 | 86.17 | 58.54 | 82.44 | 84.82 | 88.71 | 90.89 | 88.77 | 87.37 | 71.96 | 71.18 | 81.23 | 81.4 | 77.13 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CFAArchitecture=One-stage, Pre-trained=ImageNet2024.03 | 81.23 | 76.67 | 89.08 | 83.2 | 54.37 | 66.87 | 80.96 | 87.17 | 90.21 | 84.32 | 86.09 | 52.34 | 69.94 | 75.52 | 80.76 | 67.96 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CFAPretrain=IN, Backbone=ResNet-152, MS=false2022.12 | 81.23 | 76.67 | 89.08 | 83.2 | 54.37 | 66.87 | 80.96 | 87.17 | 90.21 | 84.32 | 86.09 | 52.34 | 69.94 | 75.52 | 80.76 | 67.96 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CFAModel=R152, Pre-training=IN2023.08 | 81.23 | 76.67 | 89.08 | 83.2 | 54.37 | 66.87 | 80.96 | 87.17 | 90.21 | 84.32 | 86.09 | 52.34 | 69.94 | 75.52 | 80.76 | 67.96 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CSL(YOLOv5-m)Backbone=YOLOv5-m, Two-stage detection framework=false, Multi-scale training=false2022.03 | 81.19 | — | 88.36 | 85.24 | 53.86 | 62.35 | 84.72 | 88.57 | 90.77 | 80.98 | 88.09 | 53.92 | 61.9 | 75.99 | 80.43 | 68.59 | — | — | — | — | 76.33 | — | — | — | — | — | — | — | — | — | — | — | |
| RHINOBackbone=Swin-T, Scale configuration=single-scale training and testing2023.05 | 81.13 | — | 88.18 | 84.77 | 58.54 | 77.73 | 85.58 | 89.18 | 90.86 | 86.96 | 86.38 | 65.49 | 71.29 | 78.17 | 82.8 | 64.26 | — | — | — | — | 79.42 | — | — | — | — | — | — | — | — | — | — | — | |
| OrientedFormerBackbone=Swin-T2026.02 | 81.08 | — | 88.74 | 78.94 | 53.43 | 72.05 | 84.22 | 88.4 | 90.9 | 86.23 | 86.65 | 61.05 | 63.11 | 75.78 | 73.02 | 54.62 | — | — | — | — | 75.88 | — | — | — | — | — | — | — | — | — | — | — | |
| imTED-OModel=ViT-B, Pre-training=M2023.08 | 81.07 | 81.1 | 89.41 | 84.02 | 62.38 | 81.05 | 86.12 | 88.47 | 90.72 | 87.03 | 87.18 | 68.01 | 71.84 | 78.87 | 79.92 | 80.39 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| STD-OModel=ViT-B, Pre-training=M2023.08 | 81.06 | 81.66 | 88.56 | 84.53 | 62.08 | 81.8 | 85.06 | 88.43 | 90.59 | 86.84 | 86.95 | 72.13 | 71.54 | 84.3 | 82.05 | 78.94 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Oriented FormerBackbone=R-502026.02 | 81.02 | — | 88.14 | 79.13 | 51.96 | 67.34 | 83.26 | 88.29 | 90.9 | 85.57 | 86.25 | 60.84 | 66.36 | 73.81 | 71.23 | 56.49 | — | — | — | — | 75.37 | — | — | — | — | — | — | — | — | — | — | — | |
| S2A-NetStage=Single-stage, Backbone=ResNet-502021.03 | 80.94 | 76.38 | 89.07 | 82.22 | 53.63 | 69.88 | 82.12 | 88.72 | 90.73 | 83.77 | 86.92 | 63.78 | 67.86 | 76.51 | 73.03 | 56.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| STD-OModel=HiViT-B, Pre-training=M2023.08 | 80.9 | 82.24 | 89.15 | 85.03 | 60.79 | 82.06 | 85.76 | 88.45 | 90.83 | 87.71 | 87.29 | 73.99 | 71.25 | 85.18 | 82.17 | 82.95 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ReDiffDetBackbone=R-502026.02 | 80.9 | — | 85.45 | 77.84 | 50.39 | 71.63 | 84.85 | 88.77 | 90.88 | 87 | 86.08 | 63.3 | 63.43 | 76.01 | 76.03 | 60.16 | — | — | — | — | 76.18 | — | — | — | — | — | — | — | — | — | — | — | |
| SASMPretrain=IN, Backbone=ResNeXt-101, MS=false2022.12 | 80.87 | 77.19 | 88.41 | 83.32 | 54 | 74.34 | 84.1 | 88.04 | 90.74 | 82.85 | 86.26 | 63.96 | 66.78 | 78.4 | 73.84 | 61.97 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AO2-DETRArchitecture=One-stage, Pre-trained=ImageNet, Parameters=74.3M, FLOPs=304G2024.03 | 80.86 | 79.22 | 89.95 | 84.52 | 56.9 | 74.83 | 83.47 | 88.47 | 90.87 | 86.12 | 88.55 | 63.21 | 65.09 | 79.09 | 82.88 | 73.46 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RVSA-OModel=HiViT-B, Pre-training=M2023.08 | 80.84 | 81.32 | 88.71 | 86.02 | 62.45 | 80.24 | 85.19 | 88.47 | 90.7 | 86.64 | 86.37 | 67.41 | 74.63 | 79.04 | 80.8 | 82.28 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-R-lPretrain=IN, Backbone=RTMDet-l backbone, MS=false2022.12 | 80.81 | 78.85 | 89.43 | 84.21 | 55.2 | 75.06 | 84.53 | 88.97 | 90.9 | 87.38 | 87.25 | 63.09 | 67.87 | 78.09 | 80.78 | 69.13 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-R-lBackbone=CSPNeXt-l, Scale configuration=single-scale training and testing2023.05 | 80.81 | — | 89.43 | 84.21 | 55.2 | 75.06 | 84.53 | 88.97 | 90.9 | 87.38 | 87.25 | 63.09 | 67.87 | 78.09 | 80.78 | 69.13 | — | — | — | — | 78.85 | — | — | — | — | — | — | — | — | — | — | — | |
| Oriented RepPointsPretrain=IN, Backbone=Swin-tiny, MS=false2022.12 | 80.7 | 77.63 | 89.11 | 82.32 | 56.71 | 74.95 | 83.73 | 87.67 | 90.81 | 87.11 | 85.85 | 63.6 | 68.6 | 75.95 | 73.54 | 63.76 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| EIE-Detsupervision=HBox-supervised, end-to-end=true2025.02 | 80.7 | — | 87.7 | 70.2 | 41.5 | 60.5 | 76.3 | 86.3 | 90.9 | 82.6 | 84.7 | 53.1 | 64.5 | 58.1 | 70.4 | 43.8 | — | — | — | — | 70.1 | — | — | — | — | — | — | — | — | — | — | — | |
| Oriented ReppointsBackbone=Swin-T2026.02 | 80.7 | — | 89.11 | 82.32 | 56.71 | 74.95 | 83.7 | 87.67 | 90.81 | 87.11 | 85.85 | 63.6 | 68.6 | 75.95 | 73.54 | 63.76 | — | — | — | — | 77.63 | — | — | — | — | — | — | — | — | — | — | — | |
| imTED-OModel=ViT-B, Pre-training=M†2023.08 | 80.68 | 81.01 | 87.63 | 85.23 | 61.73 | 81.11 | 85.37 | 88.26 | 90.8 | 86.38 | 87.21 | 67.93 | 69.81 | 84.06 | 81.25 | 77.76 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-R-sPretrain=IN, Backbone=RTMDet-s backbone, MS=true2022.12 | 80.62 | 79.98 | 88.16 | 86.09 | 56.8 | 78.79 | 85.06 | 88.64 | 90.82 | 86.9 | 86.7 | 66.23 | 70.22 | 78.17 | 81.71 | 74.58 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-RArchitecture=One-stage, Pre-trained=ImageNet, Parameters=52.3M, FLOPs=205G2024.03 | 80.58 | 80.54 | 88.36 | 84.96 | 57.33 | 80.46 | 84.88 | 88.08 | 90.9 | 86.32 | 87.57 | 69.29 | 70.61 | 78.63 | 80.97 | 79.24 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-R-lPretrain=IN, Backbone=RTMDet-l backbone, MS=true2022.12 | 80.58 | 80.54 | 88.36 | 84.96 | 57.33 | 80.46 | 84.88 | 88.08 | 90.9 | 86.32 | 87.57 | 69.29 | 70.61 | 78.63 | 80.97 | 79.24 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Proposed MethodBackbone=ResNet-50, multiscale=true2024.07 | 80.47 | 81.11 | 89.69 | 85.25 | 60.37 | 81.78 | 85.65 | 88.8 | 90.87 | 86.45 | 87.75 | 72.22 | 72.43 | 78.78 | 81.17 | 75.01 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| imTED-OModel=HiViT-B, Pre-training=M2023.08 | 80.44 | 81.61 | 88.24 | 84.94 | 63.13 | 81.7 | 83.3 | 88.18 | 90.78 | 88.14 | 88.23 | 71.37 | 71.83 | 78.77 | 80.84 | 84.25 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Rotated FCOSBackbone=R-50, reproduced=true, training schedule=3x2026.02 | 80.42 | — | 88.52 | 77.54 | 47.06 | 63.78 | 80.5 | 87.34 | 90.39 | 77.83 | 84.13 | 55.45 | 65.84 | 66.15 | 72.77 | 49.17 | — | — | — | — | 72.45 | — | — | — | — | — | — | — | — | — | — | — | |
| PPYOLOE-R-sPretrain=IN, Backbone=CSPRepResNet-s, MS=false2022.12 | 80.41 | 73.82 | 88.8 | 79.24 | 45.92 | 66.88 | 82.95 | 88.2 | 90.61 | 82.91 | 86.37 | 55.8 | 64.11 | 65.09 | 79.5 | 50.43 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TLFBackbone=ConvNeXt-T, Multiscale training and testing=true2024.07 | 80.37 | 82.02 | 89.59 | 86.93 | 61.91 | 82.94 | 85.9 | 88.62 | 90.85 | 87.23 | 87.83 | 72.54 | 73.92 | 79.35 | 80.4 | 81.89 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TLFBackbone=ConvNeXt-T [41], multiscale=true2024.07 | 80.37 | 82.02 | 89.59 | 86.93 | 61.91 | 82.94 | 85.9 | 88.62 | 90.85 | 87.23 | 87.83 | 72.54 | 73.92 | 79.35 | 80.4 | 81.89 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Oriented R-CNNBackbone=PKINet-S, Multiscale training and testing=false2024.07 | 80.25 | 78.39 | 89.72 | 84.2 | 55.81 | 77.63 | 84.45 | 88.12 | 90.88 | 87.57 | 86.07 | 66.86 | 70.23 | 77.47 | 73.62 | 62.94 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TLFBackbone=ConvNeXt-T, Multiscale training and testing=false2024.07 | 80.25 | 78.93 | 89.6 | 85.83 | 56.19 | 77.1 | 84.98 | 88.4 | 90.85 | 87.99 | 86.14 | 69.8 | 70.32 | 76.97 | 74.09 | 65.49 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Oriented R-CNNBackbone=PKINet-S [16], multiscale=false2024.07 | 80.25 | 78.39 | 89.72 | 84.2 | 55.81 | 77.63 | 84.45 | 88.12 | 90.88 | 87.57 | 86.07 | 66.86 | 70.23 | 77.47 | 73.62 | 62.94 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TLFBackbone=ConvNeXt-T [41], multiscale=false2024.07 | 80.25 | 78.93 | 89.6 | 85.83 | 56.19 | 77.1 | 84.98 | 88.4 | 90.85 | 87.99 | 86.14 | 69.8 | 70.32 | 76.97 | 74.09 | 65.49 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RHINOBackbone=R-50, Scale configuration=single-scale training and testing2023.05 | 80.23 | — | 88.28 | 85.19 | 55.89 | 72.75 | 83.14 | 89.07 | 90.88 | 87.12 | 86.89 | 65.37 | 71.68 | 77.75 | 81.24 | 64.71 | — | — | — | — | 78.68 | — | — | — | — | — | — | — | — | — | — | — | |
| QPDetBackbone=ResNet-50, multiscale=true2024.07 | 80.21 | 81 | 90.14 | 85.31 | 60.98 | 79.92 | 85.04 | 88.8 | 90.87 | 86.45 | 88.04 | 70.88 | 71.72 | 82.99 | 80.55 | 73.19 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ARS-DETRBackbone=Swin-T, reproduced=true, training schedule=3x2026.02 | 80.19 | — | 87.78 | 78.58 | 52.58 | 67.69 | 84.32 | 88.19 | 90.68 | 85.92 | 84.76 | 55.18 | 66.89 | 74.57 | 79.09 | 60.35 | — | — | — | — | 75.79 | — | — | — | — | — | — | — | — | — | — | — | |
| ARS-DETRBackbone=Swin-T, Scale configuration=single-scale training and testing2023.05 | 80.19 | — | 87.78 | 78.58 | 52.58 | 67.69 | 84.32 | 88.19 | 90.68 | 85.92 | 84.76 | 55.18 | 66.89 | 74.57 | 79.09 | 60.35 | — | — | — | — | 75.79 | — | — | — | — | — | — | — | — | — | — | — | |
| DBAModel=ReR50, Pre-training=IN2023.08 | 80.11 | 80.37 | 89.92 | 83.84 | 59.65 | 79.88 | 87.96 | 88.17 | 90.31 | 88.93 | 88.46 | 68.93 | 65.94 | 78.04 | 79.69 | 75.78 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| KFloUArchitecture=Two-stage, Pre-trained=ImageNet, Parameters=58.8M, FLOPs=206G2024.03 | 80.1 | 80.93 | 89.44 | 84.41 | 62.22 | 82.51 | 86.07 | 88.68 | 90.9 | 87.32 | 88.38 | 72.8 | 71.95 | 78.96 | 74.95 | 75.27 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RoI Trans. + KFIoUPretrain=IN, Backbone=Swin-tiny, MS=false2022.12 | 80.1 | 80.93 | 89.44 | 84.41 | 62.22 | 82.51 | 86.07 | 88.68 | 90.9 | 87.32 | 88.38 | 72.8 | 71.95 | 78.96 | 74.95 | 75.27 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| KFIoUModel=Swin-T, Pre-training=IN2023.08 | 80.1 | 80.93 | 89.44 | 84.41 | 62.22 | 82.51 | 86.07 | 88.68 | 90.9 | 87.32 | 88.38 | 72.8 | 71.95 | 78.96 | 74.95 | 75.27 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Proposed MethodBackbone=ResNet-502024.07 | 80.09 | 77.19 | 89.54 | 83.14 | 55.32 | 71.56 | 83.58 | 88.2 | 90.9 | 87.93 | 85.77 | 65.69 | 66.3 | 74.8 | 71.29 | 63.72 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RTMDet-R-mPretrain=IN, Backbone=RTMDet-m backbone, MS=true2022.12 | 80.04 | 80.26 | 87.1 | 85.83 | 56.3 | 80.28 | 84.67 | 88.22 | 90.88 | 88.49 | 87.57 | 70.74 | 69.99 | 78.35 | 80.88 | 74.53 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CFABackbone=ResNet1012026.02 | 80 | — | 89.3 | 81.7 | 51.8 | 67.2 | 78.3 | 84.5 | 90.8 | 83.4 | 85.5 | 54.9 | 67.8 | 73 | 70.2 | 65 | — | — | — | — | 75.1 | — | — | — | — | — | — | — | — | — | — | — | |
| Oriented R-CNNBackbone=ViTAE-B + RVSA(MAE), Scale configuration=single-scale training and testing2023.05 | 79.99 | — | 89.38 | 84.26 | 59.39 | 73.19 | 85.36 | 88.08 | 90.87 | 88.5 | 86.53 | 58.93 | 72.24 | 77.31 | 79.59 | 71.24 | — | — | — | — | 78.99 | — | — | — | — | — | — | — | — | — | — | — | |
| RVSAArchitecture=Two-stage, Pre-trained=MA, Parameters=114.4M, FLOPs=414G2024.03 | 79.98 | 81.24 | 88.97 | 85.76 | 61.46 | 81.27 | 85.31 | 88.3 | 90.84 | 85.06 | 87.5 | 66.77 | 73.11 | 84.75 | 81.88 | 77.58 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RVSA-OModel=ViTAE-B, Pre-training=M2023.08 | 79.98 | 81.24 | 88.97 | 85.76 | 61.46 | 81.27 | 85.31 | 88.3 | 90.84 | 85.06 | 87.5 | 66.77 | 73.11 | 84.75 | 81.88 | 77.58 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PSD-SQBackbone=R-50, Scale configuration=single-scale training and testing2023.05 | 79.98 | — | 89.67 | 85.43 | 57.28 | 75.24 | 81.21 | 88.3 | 89.88 | 88.05 | 86.35 | 69.63 | 68.45 | 75.33 | 69.46 | 72.65 | — | — | — | — | 78.46 | — | — | — | — | — | — | — | — | — | — | — | |
| KFIoUModel=Swin-B, Pre-training=IN2023.08 | 79.97 | 81.18 | 89.26 | 86.48 | 62.09 | 82.86 | 85.64 | 88.47 | 90.7 | 86.69 | 87.54 | 71.84 | 68.74 | 79.62 | 81.11 | 76.64 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CFABackbone=R-50, reproduced=true, training schedule=3x2026.02 | 79.95 | — | 88.34 | 83.09 | 51.92 | 72.23 | 78.68 | 87.25 | 90.9 | 85.38 | 85.71 | 59.63 | 63.05 | 73.33 | 70.36 | 47.86 | — | — | — | — | 74.51 | — | — | — | — | — | — | — | — | — | — | — | |
| S2ANetArchitecture=One-stage, Pre-trained=ImageNet2024.03 | 79.94 | 79.42 | 88.89 | 83.6 | 57.74 | 81.95 | 83.19 | 89.11 | 90.78 | 84.87 | 87.81 | 70.3 | 68.25 | 78.3 | 77.01 | 69.58 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| S2A-Netbackbone=R50-FPN, scale=multi-scale2021.03 | 79.94 | 79.42 | 88.89 | 83.6 | 57.74 | 81.95 | 83.19 | 89.11 | 90.78 | 84.87 | 87.81 | 70.3 | 68.25 | 78.3 | 77.01 | 69.58 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| S2ANetPretrain=IN, Backbone=ResNet-50, MS=false2022.12 | 79.94 | 79.42 | 88.89 | 83.6 | 57.74 | 81.95 | 83.19 | 89.11 | 90.78 | 84.87 | 87.81 | 70.3 | 68.25 | 78.3 | 77.01 | 69.58 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| S2ANetModel=R50, Pre-training=IN2023.08 | 79.94 | 79.42 | 88.89 | 83.6 | 57.74 | 81.95 | 83.19 | 89.11 | 90.78 | 84.87 | 87.81 | 70.3 | 68.25 | 78.3 | 77.01 | 69.58 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| KLDBackbone=R-50, reproduced=true, training schedule=3x2026.02 | 79.85 | — | 89.08 | 84.18 | 43.77 | 72.33 | 73.58 | 85.69 | 90.88 | 85.14 | 81.96 | 65.86 | 64.6 | 63.6 | 68.26 | 53.19 | — | — | — | — | 73.46 | — | — | — | — | — | — | — | — | — | — | — | |
| S2A-Net + RILStage=Single-stage, Backbone=ResNet-50, Module=Representation Invariance Loss2021.03 | 79.81 | 77.62 | 89.31 | 80.77 | 54.07 | 76.38 | 81.99 | 89.13 | 90.72 | 83.58 | 87.22 | 64.42 | 67.56 | 78.08 | 79.17 | 62.07 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| H2RBox-v2Supervision=HBox-supervised2023.11 | 79.8 | — | 89 | 74.4 | 51 | 60.5 | 75.3 | 86.9 | 90.9 | 86.1 | 85 | 59.2 | 63.2 | 65.2 | 71.6 | 49.7 | — | — | — | — | 72.52 | — | 75.88 | — | — | — | — | — | — | — | — | — | |
| H2RBox-v2supervision=HBox-supervised2023.11 | 79.8 | — | 89 | 74.4 | 51 | 60.5 | 75.3 | 86.9 | 90.9 | 86.1 | 85 | 59.2 | 63.2 | 65.2 | 71.6 | 49.7 | — | — | — | — | 72.52 | — | 75.88 | — | — | — | — | — | — | — | — | — | |
| FCOSSupervision=RBox-supervised2023.11 | 79.8 | — | 89.1 | 76.9 | 50.1 | 63.2 | 79.8 | 87.1 | 90.4 | 80.8 | 84.6 | 59.7 | 66.3 | 65.8 | 71.3 | 41.7 | — | — | — | — | 72.44 | — | — | — | — | — | — | — | — | — | — | — | |
| H2RBox-v2Supervision=HBox-to-RBox2023.11 | 79.8 | — | 89 | 74.4 | 50 | 60.5 | 75.3 | 86.9 | 90.9 | 85.1 | 85 | 59.2 | 63.2 | 65.2 | 70.5 | 49.7 | — | — | — | — | 72.31 | — | — | — | — | — | — | — | — | — | — | — | |
| FCOSsupervision=RBox-supervised, end-to-end=true2025.02 | 79.8 | — | 89.1 | 76.9 | 50.1 | 63.2 | 79.8 | 87.1 | 90.4 | 80.8 | 84.6 | 59.7 | 66.3 | 65.8 | 71.3 | 41.7 | — | — | — | — | 72.44 | — | — | — | — | — | — | — | — | — | — | — | |
| H2RBox-v2supervision=HBox-supervised, end-to-end=true2025.02 | 79.8 | — | 89 | 74.4 | 50 | 60.5 | 75.3 | 86.9 | 90.9 | 85.1 | 85 | 59.2 | 63.2 | 65.2 | 70.5 | 49.7 | — | — | — | — | 72.31 | — | — | — | — | — | — | — | — | — | — | — | |
| SASMArchitecture=One-stage, Pre-trained=ImageNet2024.03 | 79.78 | 79.17 | 89.54 | 85.94 | 57.73 | 78.41 | 84.19 | 89.25 | 90.87 | 58.8 | 87.27 | 63.82 | 67.81 | 78.67 | 79.35 | 69.37 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SASMModel=RX101, Pre-training=IN2023.08 | 79.78 | 79.17 | 89.54 | 85.94 | 57.73 | 78.41 | 84.19 | 89.25 | 90.87 | 58.8 | 87.27 | 63.82 | 67.81 | 78.67 | 79.35 | 69.37 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GWDBackbone=R-50, reproduced=true, training schedule=3x2026.02 | 79.76 | — | 89.34 | 80.07 | 41.94 | 72.58 | 69.93 | 85.38 | 90.76 | 83.16 | 82.38 | 62.75 | 66.06 | 60.82 | 69.39 | 57.84 | — | — | — | — | 72.81 | — | — | — | — | — | — | — | — | — | — | — | |
| O-RCNNArchitecture=Two-stage, Pre-trained=ImageNet, Parameters=41.1M, FLOPs=199G2024.03 | 79.71 | 80.87 | 89.84 | 85.43 | 61.09 | 79.82 | 85.35 | 88.82 | 90.88 | 86.68 | 87.73 | 72.21 | 70.8 | 82.42 | 78.18 | 74.11 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Oriented RCNNPretrain=IN, Backbone=ResNet-50, MS=false2022.12 | 79.71 | 80.87 | 89.84 | 85.43 | 61.09 | 79.82 | 85.35 | 88.82 | 90.88 | 86.68 | 87.73 | 72.21 | 70.8 | 82.42 | 78.18 | 74.11 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| O-RCNNModel=R50, Pre-training=IN2023.08 | 79.71 | 80.87 | 89.84 | 85.43 | 61.09 | 79.82 | 85.35 | 88.82 | 90.88 | 86.68 | 87.73 | 72.21 | 70.8 | 82.42 | 78.18 | 74.11 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Oriented R-CNNBackbone=ResNet-50, multiscale=true2024.07 | 79.71 | 80.87 | 89.84 | 85.43 | 61.09 | 79.82 | 85.35 | 88.82 | 90.88 | 86.68 | 87.73 | 72.21 | 70.08 | 82.42 | 78.18 | 74.11 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Oriented ReppointsBackbone=R-50, reproduced=true, training schedule=3x2026.02 | 79.61 | — | 88.52 | 80.62 | 52.68 | 73.04 | 80.73 | 87.76 | 90.89 | 81.82 | 85.33 | 59.95 | 64.88 | 73.81 | 69.84 | 46.18 | — | — | — | — | 74.38 | — | — | — | — | — | — | — | — | — | — | — | |
| Point2RBox-v2supervision=Point-supervised, end-to-end=false2025.02 | 79.6 | — | 88 | 72.6 | 8 | 46.2 | 76.3 | 86.9 | 89.1 | 79.7 | 82.9 | 26.2 | 45.3 | 45.8 | 66.3 | 46.3 | — | — | — | — | 62.61 | — | — | — | — | — | — | — | — | — | — | — |