Object Detection on DOTA
79.42mAPS²A-Net*
Evaluation results
| Method | Links | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| S²A-Net*Detection stage=one-stage, Backbone=R-50-FPN, Multi-scale training and testing=true2020.08 | 79.42 | 88.89 | 83.6 | 57.74 | 81.95 | 79.94 | 83.19 | 89.11 | 90.78 | 84.87 | 87.81 | 68.25 | 78.3 | 77.01 | 69.58 | 70.3 | 16 | — | — | — | |
| S²A-Net*Detection stage=one-stage, Backbone=R-101-FPN, Multi-scale training and testing=true2020.08 | 79.15 | 89.28 | 84.11 | 56.95 | 79.21 | 80.18 | 82.93 | 89.21 | 90.86 | 84.66 | 87.61 | 68.23 | 78.58 | 78.2 | 65.55 | 71.66 | 12.7 | — | — | — | |
| S²A-NetDetection stage=one-stage, Backbone=R-101-FPN2020.08 | 76.11 | 88.7 | 81.41 | 54.28 | 69.75 | 78.04 | 80.54 | 88.04 | 90.69 | 84.75 | 86.22 | 65.81 | 76.16 | 73.37 | 58.86 | 65.03 | 12.7 | — | — | — | |
| CenterMap-NetDetection stage=two-stage, Backbone=R-101-FPN, Multi-scale training and testing=true2020.08 | 76.03 | 89.83 | 84.41 | 54.6 | 70.25 | 77.66 | 78.32 | 87.19 | 90.66 | 84.89 | 85.27 | 69.23 | 74.13 | 71.56 | 66.06 | 56.46 | — | — | — | — | |
| Xu et al.Detection stage=two-stage, Backbone=R-101-FPN2020.08 | 75.02 | 89.64 | 85 | 52.26 | 77.34 | 73.01 | 73.14 | 86.82 | 90.74 | 79.02 | 86.81 | 70.91 | 72.94 | 70.86 | 57.32 | 59.55 | 10 | — | — | — | |
| S²A-NetDetection stage=one-stage, Backbone=R-50-FPN2020.08 | 74.12 | 89.11 | 82.84 | 48.37 | 71.11 | 78.11 | 78.39 | 87.25 | 90.83 | 84.9 | 85.64 | 62.6 | 65.26 | 69.13 | 57.94 | 60.36 | 16 | — | — | — | |
| S²A-Net+Detection stage=one-stage, Backbone=R-50-FPN, Input image type=original images2020.08 | 74.01 | 89.11 | 81.51 | 48.75 | 72.85 | 78.23 | 76.77 | 86.95 | 90.84 | 83.59 | 85.52 | 61.63 | 66.55 | 68.94 | 56.24 | 62.7 | 22.6 | — | — | — | |
| R³DetDetection stage=one-stage, Backbone=R-152-FPN2020.08 | 73.74 | 89.49 | 81.17 | 50.53 | 66.1 | 70.92 | 78.66 | 78.21 | 90.81 | 85.26 | 84.23 | 63.77 | 68.16 | 69.83 | 67.17 | 61.81 | — | — | — | — | |
| SCRDetDetection stage=two-stage, Backbone=R-101-FPN2020.08 | 72.61 | 89.98 | 80.65 | 52.09 | 68.36 | 68.36 | 60.32 | 72.41 | 90.85 | 87.94 | 86.86 | 66.68 | 66.25 | 68.24 | 65.21 | 65.02 | — | — | — | — | |
| CenterMap-NetDetection stage=two-stage, Backbone=R-50-FPN2020.08 | 71.74 | 88.88 | 81.24 | 53.15 | 60.65 | 78.62 | 66.55 | 78.1 | 88.83 | 77.8 | 83.61 | 66.19 | 72.1 | 72.36 | 58.7 | 49.36 | — | — | — | — | |
| R³DetDetection stage=one-stage, Backbone=R-101-FPN2020.08 | 71.69 | 89.54 | 81.99 | 48.46 | 62.52 | 70.48 | 74.29 | 77.54 | 90.8 | 81.39 | 83.54 | 59.82 | 65.44 | 67.46 | 60.05 | 61.97 | — | — | — | — | |
| DRNDetection stage=one-stage, Backbone=H-1042020.08 | 70.7 | 88.91 | 80.22 | 43.52 | 63.35 | 73.48 | 70.69 | 84.94 | 90.14 | 83.85 | 84.11 | 58.41 | 67.62 | 68.6 | 52.5 | 50.12 | — | — | — | — | |
| RMK RetinaNetBackbone=ResNet-502026.03 | 70.38 | 89.9 | 79.9 | 45.7 | 78.01 | 69.89 | 74.91 | 74.83 | 91.63 | 78.5 | 83.14 | 54.93 | 64.27 | 60.53 | 49.49 | 60.08 | — | — | — | — | |
| CADNetDetection stage=two-stage, Backbone=R-101-FPN2020.08 | 69.9 | 87.8 | 82.4 | 49.4 | 73.5 | 71.1 | 63.5 | 76.6 | 90.9 | 79.2 | 73.3 | 60.9 | 62 | 67 | 62.2 | 48.4 | — | — | — | — | |
| CADNetBackbone=ResNet-1012026.03 | 69.9 | 87.8 | 82.4 | 49.4 | 73.5 | 71.1 | 63.5 | 76.6 | 90.9 | 79.2 | 73.3 | 60.9 | 62 | 67 | 62.2 | 48.4 | — | — | — | — | |
| RoI TransformerFPN=true2018.12 | 69.56 | 88.64 | 78.52 | 43.44 | 75.92 | 68.81 | 73.68 | 83.59 | 90.74 | 77.27 | 81.46 | 53.54 | 62.83 | 58.93 | 47.67 | — | — | — | — | — | |
| RoI Trans.Detection stage=two-stage, Backbone=R-101-FPN, Multi-scale training and testing=true2020.08 | 69.56 | 88.64 | 78.52 | 43.44 | 75.92 | 68.81 | 73.68 | 83.59 | 90.74 | 77.27 | 81.46 | 53.54 | 62.83 | 58.93 | 47.67 | 58.39 | 5.9 | — | — | — | |
| RoI Trans.Backbone=ResNet-1012026.03 | 69.56 | 88.64 | 78.52 | 43.44 | 75.92 | 68.81 | 73.68 | 90.74 | 77.27 | 81.46 | 86.15 | 53.54 | 62.83 | 58.93 | 47.67 | 58.39 | — | — | — | — | |
| RetinaNetDetection stage=one-stage, Backbone=R-101-FPN2020.08 | 68.72 | 88.82 | 81.74 | 44.44 | 67.11 | 65.71 | 55.37 | 72.77 | 90.55 | 82.83 | 76.3 | 63.64 | 63.71 | 69.73 | 53.37 | 54.19 | 12.7 | — | — | — | |
| Rotation RetinaNetBackbone=ResNet-502026.03 | 68.49 | 88.01 | 78.01 | 43.81 | 76.12 | 68 | 73.05 | 72.91 | 89.74 | 76.61 | 81.25 | 53.04 | 62.38 | 58.64 | 47.6 | 58.19 | — | — | — | — | |
| Azimi et al.Detection stage=two-stage, Backbone=R-101-FPN2020.08 | 68.16 | 81.36 | 74.3 | 47.7 | 70.32 | 64.89 | 67.82 | 69.98 | 90.76 | 79.06 | 78.2 | 62.9 | 67.02 | 64.17 | 50.23 | 53.64 | — | — | — | — | |
| Azimi et al.Backbone=ResNet-1012026.03 | 68.16 | 81.36 | 74.3 | 47.7 | 70.32 | 64.89 | 67.82 | 69.98 | 90.76 | 79.06 | 78.2 | 62.9 | 67.02 | 64.17 | 50.23 | 53.64 | — | — | — | — | |
| RoI TransformerFPN=false2018.12 | 67.74 | 88.53 | 77.91 | 37.63 | 74.08 | 66.53 | 62.97 | 66.57 | 90.5 | 79.46 | 76.75 | 56.73 | 62.54 | 61.29 | 55.56 | — | — | — | — | — | |
| BaselineFPN=true2018.12 | 66.95 | 88.02 | 76.99 | 36.7 | 72.54 | 70.15 | 61.79 | 75.77 | 90.14 | 73.81 | 85.04 | 62.63 | 53.3 | 59.54 | 41.91 | — | — | — | — | — | |
| Deformable R-FCNBackbone=ResNet-1012026.03 | 66.56 | 87.97 | 76.69 | 46.97 | 68.76 | 55.86 | 63.5 | 56.9 | 90.13 | 75.81 | 64.8 | 60.01 | 74.96 | 71.41 | 52.95 | 51.73 | — | — | — | — | |
| mtLoRArank=162026.03 | 64 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 91.7 | |
| DPSRPFPN=false2018.12 | 63.89 | 81.18 | 77.42 | 35.48 | 70.41 | 56.74 | 50.42 | 53.56 | 89.97 | 79.68 | 76.48 | 59.94 | 53.34 | 64.04 | 47.76 | — | — | — | — | — | |
| Yang et al.FPN=true2018.12 | 62.29 | 81.25 | 71.41 | 36.53 | 67.44 | 61.16 | 50.91 | 56.6 | 90.67 | 68.09 | 72.39 | 55.6 | 62.44 | 53.35 | 51.47 | — | — | — | — | — | |
| HydraLoRArank=16, hyperparameter search=optimal, BLC optimization=true2026.03 | 61.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 89.1 | |
| LoRAHubrank=162026.03 | 61.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 88.9 | |
| MMoELoRArank=162026.03 | 61.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 89.4 | |
| RRPNFPN=false2018.12 | 61.01 | 80.94 | 65.75 | 35.34 | 67.44 | 59.92 | 50.91 | 55.81 | 90.67 | 66.92 | 72.39 | 52.23 | 55.14 | 53.35 | 48.22 | — | — | — | — | — | |
| RRPNBackbone=ResNet-1012026.03 | 61.01 | 88.52 | 71.2 | 31.66 | 59.3 | 51.85 | 56.19 | 57.25 | 90.81 | 72.84 | 67.38 | 52.84 | 53.08 | 51.94 | 53.58 | 56.69 | — | — | — | — | |
| Earth-Agent2025.09 | 60.88 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| R2CNNFPN=false2018.12 | 60.67 | 88.52 | 71.2 | 31.66 | 59.3 | 51.85 | 56.19 | 57.25 | 90.81 | 72.84 | 67.38 | 52.84 | 53.08 | 51.94 | 53.58 | — | — | — | — | — | |
| R2CNNBackbone=ResNet-1012026.03 | 60.67 | 80.94 | 65.67 | 35.34 | 67.44 | 59.92 | 50.91 | 55.81 | 90.67 | 66.92 | 72.39 | 52.23 | 55.14 | 53.35 | 48.22 | 55.06 | — | — | — | — | |
| PIoUBackbone=DLA-342026.03 | 60.5 | 80.9 | 69.7 | 24.1 | 60.2 | 38.3 | 64.4 | 64.8 | 90.9 | 77.2 | 70.4 | 37.1 | 57.1 | 61.9 | 64 | 46.5 | — | — | — | — | |
| BaselineFPN=false2018.12 | 58.31 | 81.06 | 76.81 | 27.22 | 69.75 | 38.99 | 39.07 | 38.3 | 89.97 | 75.53 | 65.74 | 59.37 | 48.11 | 56.86 | 44.46 | — | — | — | — | — | |
| R-DFPNFPN=true2018.12 | 57.94 | 80.92 | 65.82 | 33.77 | 58.94 | 55.77 | 50.94 | 54.78 | 90.33 | 66.34 | 68.66 | 51.76 | 55.1 | 51.32 | 35.88 | — | — | — | — | — | |
| R-DFPNBackbone=ResNet-1012026.03 | 57.94 | 80.92 | 65.82 | 33.77 | 58.94 | 55.77 | 50.94 | 54.78 | 90.33 | 66.34 | 68.66 | 51.76 | 55.1 | 51.32 | 35.88 | 48.73 | — | — | — | — | |
| IENetBackbone=ResNet-1012026.03 | 57.14 | 80.2 | 62.54 | 39.82 | 32.07 | 49.71 | 65.01 | 52.58 | 81.45 | 44.66 | 78.51 | 56.73 | 64.4 | 64.24 | 36.75 | 46.54 | — | — | — | — | |
| FR-OFPN=false2018.12 | 54.13 | 79.42 | 77.13 | 17.7 | 64.05 | 35.3 | 38.02 | 37.16 | 89.41 | 69.64 | 59.28 | 52.91 | 47.89 | 47.4 | 46.3 | — | — | — | — | — | |
| FR-ODetection stage=two-stage, Backbone=R-1012020.08 | 54.13 | 79.42 | 77.13 | 17.7 | 64.05 | 35.3 | 38.02 | 37.16 | 89.41 | 69.64 | 59.28 | 52.91 | 47.89 | 47.4 | 46.3 | 50.3 | — | — | — | — | |
| R-FCNBackbone=ResNet-1012026.03 | 52.58 | 81.01 | 58.96 | 31.64 | 58.97 | 49.77 | 45.04 | 49.29 | 68.99 | 52.07 | 67.42 | 51.44 | 45.15 | 53.3 | 33.89 | 41.83 | — | — | — | — | |
| LLaVA-1.52025.09 | 17.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LHRS-Bot2025.09 | 17.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Geochat2025.09 | 16.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Sphinx2025.09 | 15.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MiniGPT-v22025.09 | 14.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DRNDetection stage=one-stage, Backbone=H-104, Multi-scale training and testing=true2020.08 | — | 89.71 | 82.34 | 47.22 | 65.72 | 74.43 | 85.84 | 90.57 | 86.18 | 84.89 | 57.65 | 69.3 | 69.63 | 58.48 | 73.23 | 61.93 | — | — | — | — | |
| GPV-1Multi-scale inference=true2021.11 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 55 | 9.33 | — | |
| MAVLMulti-scale inference=true2021.11 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 286 | 24.2 | — | |
| MDETRMulti-scale inference=true2021.11 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 194 | 21.8 | — | |
| RetinaNetMulti-scale inference=true2021.11 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 72 | 15.6 | — |