Multi-label Image Classification on VOC 2012 (test)
96.6mAPQ2L-TResL
Evaluation results
| Method | Links | ||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Q2L-TResLResolution=448x4482021.07 | 96.6 | 99.9 | 98.2 | 99.3 | 98.1 | 90.4 | 97.7 | 97.4 | 99.4 | 92.7 | 98.7 | 89.9 | 99.4 | 99.5 | 99 | 99.4 | 88.4 | 98.8 | 89.3 | 99.6 | 96.8 | — | |
| ResNet-cut + CSRABackbone=ResNet-101 (pretrained with CutMix), extra data (MS-COCO)=true2021.08 | 96.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ADD-GCN2020.12 | 95.5 | 99.8 | 97.1 | 98.6 | 96.8 | 89.4 | 97.1 | 96.5 | 99.3 | 89 | 97.7 | 87.5 | 99.2 | 99.1 | 97.7 | 99.1 | 86.3 | 98.8 | 87 | 99.3 | 95.4 | — | |
| ADD-GCN2020.12 | 95.5 | 99.8 | 97.1 | 98.6 | 96.8 | 89.4 | 97.1 | 96.5 | 99.3 | 89 | 97.7 | 87.5 | 99.2 | 99.1 | 97.7 | 99.1 | 86.3 | 98.8 | 87 | 99.3 | 95.4 | — | |
| ADD-GCNResolution=576x5762021.07 | 95.5 | 99.8 | 97.1 | 98.6 | 96.8 | 89.4 | 97.1 | 96.5 | 99.3 | 89 | 97.7 | 87.5 | 99.2 | 99.1 | 97.7 | 99.1 | 86.3 | 98.8 | 87 | 99.3 | 95.4 | — | |
| SSGRLCOCO Pre-training=true, Two-scale Fusion=true, Model Aggregation=true2019.08 | 95.4 | 99.9 | 96.6 | 98.4 | 97 | 88.6 | 96.4 | 95.9 | 99.2 | 89 | 97.9 | 88.6 | 99.4 | 99.3 | 97.9 | 99.2 | 85.8 | 98.6 | 86.7 | 99.4 | 95.1 | — | |
| ResNet-101 + CSRABackbone=ResNet-101, extra data (MS-COCO)=true2021.08 | 95.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SSGRLCOCO Pre-training=true, Model type=Single model2019.08 | 94.8 | 99.7 | 96.1 | 97.7 | 96.5 | 86.9 | 95.8 | 95 | 98.9 | 88.3 | 97.6 | 87.4 | 99.1 | 99.2 | 97.3 | 99 | 84.8 | 98.3 | 85.8 | 99.2 | 94.1 | — | |
| SSGRL2020.12 | 94.8 | 99.7 | 96.1 | 97.7 | 96.5 | 86.9 | 95.8 | 95 | 98.9 | 88.3 | 97.6 | 87.4 | 99.1 | 99.2 | 97.3 | 99 | 84.8 | 98.3 | 85.8 | 99.2 | 94.1 | — | |
| SSGRL2020.12 | 94.8 | 99.7 | 96.1 | 97.7 | 96.5 | 86.9 | 95.8 | 95 | 98.9 | 88.3 | 97.6 | 87.4 | 99.1 | 99.2 | 97.3 | 99 | 84.8 | 98.3 | 85.8 | 99.2 | 94.1 | — | |
| SSGRLResolution=576x5762021.07 | 94.8 | 99.7 | 96.1 | 97.7 | 96.5 | 86.9 | 95.8 | 95 | 98.9 | 88.3 | 97.6 | 87.4 | 99.1 | 99.2 | 97.3 | 99 | 84.8 | 98.3 | 85.8 | 99.2 | 94.1 | — | |
| SSGRLlarger input image resolution=true, extra data (MS-COCO)=true2021.08 | 94.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ResNet-cut + CSRABackbone=ResNet-101 (pretrained with CutMix), extra data (MS-COCO)=false2021.08 | 94.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RCP+AGSModel Aggregation=true2019.08 | 94.3 | 99.8 | 94.5 | 98.1 | 96.1 | 85.5 | 96.1 | 95.5 | 99 | 90.2 | 95 | 87.8 | 98.7 | 98.4 | 97.5 | 99 | 80.1 | 95.9 | 86.5 | 98.8 | 94.6 | — | |
| MCARResolution=448x4482021.07 | 94.3 | 99.6 | 97.1 | 98.3 | 96.6 | 87 | 95.5 | 94.4 | 98.8 | 87 | 96.9 | 85 | 98.7 | 98.3 | 97.3 | 99 | 83.8 | 96.8 | 83.7 | 98.3 | 93.5 | — | |
| ResNet-101 + CSRABackbone=ResNet-101, extra data (MS-COCO)=false2021.08 | 94.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SSGRLCOCO Pre-training=false, Model type=Single model2019.08 | 93.9 | 99.5 | 95.1 | 97.4 | 96.4 | 85.8 | 94.5 | 93.7 | 98.9 | 86.7 | 96.3 | 84.6 | 98.9 | 98.6 | 96.2 | 98.7 | 82.2 | 98.2 | 84.2 | 98.1 | 93.5 | — | |
| SSGRLlarger input image resolution=true, extra data (MS-COCO)=false2021.08 | 93.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HCP+AGSModel Aggregation=true2019.08 | 93.2 | 99.8 | 94.8 | 97.7 | 95.4 | 81.3 | 96 | 94.5 | 98.9 | 88.5 | 94.1 | 86 | 98.1 | 98.3 | 97.3 | 97.3 | 76.1 | 93.9 | 84.2 | 98.2 | 92.7 | — | |
| RCPModel type=Single model2019.08 | 92.2 | 99.3 | 92.2 | 97.5 | 94.9 | 82.3 | 94.1 | 92.4 | 98.5 | 83.8 | 93.5 | 83.1 | 98.1 | 97.3 | 96 | 98.8 | 77.7 | 95.1 | 79.4 | 97.7 | 92.4 | — | |
| RCP2020.12 | 92.2 | 99.3 | 92.2 | 97.5 | 94.9 | 82.3 | 94.1 | 92.4 | 98.5 | 83.8 | 93.5 | 83.1 | 98.1 | 97.3 | 96 | 98.8 | 77.7 | 95.1 | 79.4 | 97.7 | 92.4 | — | |
| RCP2020.12 | 92.2 | 99.3 | 92.2 | 97.5 | 94.9 | 82.3 | 94.1 | 92.4 | 98.5 | 83.8 | 93.5 | 83.1 | 98.1 | 97.3 | 96 | 98.8 | 77.7 | 95.1 | 79.4 | 97.7 | 92.4 | — | |
| RCPextra data (MS-COCO)=false2021.08 | 92.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SCINetSetup=SPLC [9]2025.07 | 91.76 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SCINetSetup=LargeLoss [44]2025.07 | 90.97 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| FusionDescription=fuse scores of VeryDeep and FeV+LV-20-VD2015.04 | 90.7 | 98.9 | 93.1 | 96 | 94.1 | 76.4 | 93.5 | 90.8 | 97.9 | 78.2 | 92.1 | 82.4 | 97.2 | 96.8 | 95.7 | 98.1 | 73.9 | 93.6 | 76.8 | 97.5 | 89 | — | |
| FeV+LV (fusion)Model Aggregation=true2019.08 | 90.7 | 98.9 | 93.1 | 96 | 94.1 | 76.4 | 93.5 | 90.8 | 97.9 | 80.2 | 92.1 | 82.4 | 97.2 | 96.8 | 95.7 | 98.1 | 73.9 | 93.6 | 76.8 | 97.5 | 89 | — | |
| SCPNetSetup=SPLC [9]2025.07 | 90.55 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SCPNetSetup=LargeLoss [44]2025.07 | 90.52 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HCPModel type=Single model2019.08 | 90.5 | 99.1 | 92.8 | 97.4 | 94.4 | 79.9 | 93.6 | 89.8 | 98.2 | 78.2 | 94.9 | 79.8 | 97.8 | 97 | 93.8 | 96.4 | 74.3 | 94.7 | 71.9 | 96.7 | 88.6 | — | |
| HCP2020.12 | 90.5 | 99.1 | 92.8 | 97.4 | 94.4 | 79.9 | 93.6 | 89.8 | 98.2 | 78.2 | 94.9 | 79.8 | 97.8 | 97 | 93.8 | 96.4 | 74.3 | 94.7 | 71.9 | 96.7 | 88.6 | — | |
| HCP2020.12 | 90.5 | 99.1 | 92.8 | 97.4 | 94.4 | 79.9 | 93.6 | 89.8 | 98.2 | 78.2 | 94.9 | 79.8 | 97.8 | 97 | 93.8 | 96.4 | 74.3 | 94.7 | 71.9 | 96.7 | 88.6 | — | |
| HCPResolution=448x4482021.07 | 90.5 | 99.1 | 92.8 | 97.4 | 94.4 | 79.9 | 93.6 | 89.8 | 98.2 | 78.2 | 94.9 | 79.8 | 97.8 | 97 | 93.8 | 96.4 | 74.3 | 94.7 | 71.9 | 96.7 | 88.6 | — | |
| HCPextra data (MS-COCO)=false2021.08 | 90.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| NUS-HCP-AGSArchitecture=CNN based2015.04 | 90.3 | 99 | 91.8 | 94.8 | 92.4 | 72.6 | 95 | 91.8 | 97.4 | 85.2 | 92.9 | 83.1 | 96 | 96.6 | 96.1 | 94.9 | 68.4 | 92 | 79.6 | 97.3 | 88.5 | — | |
| FeV+LV-20-VDArchitecture=very-deep 16-layer CNN2015.04 | 89.4 | 98.4 | 92.8 | 93.4 | 90.7 | 74.9 | 93.2 | 90.2 | 96.1 | 78.2 | 89.8 | 80.6 | 95.7 | 96.1 | 95.3 | 97.5 | 73.1 | 91.2 | 75.4 | 97 | 88.2 | — | |
| FeV+LVModel type=Single model2019.08 | 89.4 | 98.4 | 92.8 | 93.4 | 90.7 | 74.9 | 93.2 | 90.2 | 96.1 | 78.2 | 89.8 | 80.6 | 95.7 | 96.1 | 95.3 | 97.5 | 73.1 | 91.2 | 75.4 | 97 | 88.2 | — | |
| FeV+LV2020.12 | 89.4 | 98.4 | 92.8 | 93.4 | 90.7 | 74.9 | 93.2 | 90.2 | 96.1 | 78.2 | 89.8 | 80.6 | 95.7 | 96.1 | 95.3 | 97.5 | 73.1 | 91.2 | 75.4 | 97 | 88.2 | — | |
| FeV+LV2020.12 | 89.4 | 98.4 | 92.8 | 93.4 | 90.7 | 74.9 | 93.2 | 90.2 | 96.1 | 78.2 | 89.8 | 80.6 | 95.7 | 96.1 | 95.3 | 97.5 | 73.1 | 91.2 | 75.4 | 97 | 88.2 | — | |
| Fev+LvResolution=448x4482021.07 | 89.4 | 98.4 | 92.8 | 93.4 | 90.7 | 74.9 | 93.2 | 90.2 | 96.1 | 78.2 | 89.8 | 80.6 | 95.7 | 96.1 | 95.3 | 97.5 | 73.1 | 91.2 | 75.4 | 97 | 88.2 | — | |
| Fev+Lvextra data (MS-COCO)=false2021.08 | 89.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VeryDeepArchitecture=very-deep CNN2015.04 | 89.3 | 99 | 89.1 | 96 | 94.1 | 74.1 | 92.2 | 85.3 | 97.9 | 79.9 | 92 | 83.7 | 97.5 | 96.5 | 94.7 | 97.1 | 63.7 | 93.6 | 75.2 | 97.4 | 87.8 | — | |
| VGG16&19+SVMModel Aggregation=true2019.08 | 89.3 | 99.1 | 89.1 | 96 | 94.1 | 74.1 | 92.2 | 85.3 | 97.9 | 79.9 | 92 | 83.7 | 97.5 | 96.5 | 94.7 | 97.1 | 63.7 | 93.6 | 75.2 | 97.4 | 87.8 | — | |
| VGG+SVMResolution=448x4482021.07 | 89.3 | 99 | 89.1 | 96 | 94.1 | 74.1 | 92.2 | 85.3 | 97.9 | 79.9 | 92 | 83.7 | 97.5 | 96.5 | 94.7 | 97.1 | 63.7 | 93.6 | 75.2 | 97.4 | 87.8 | — | |
| VGG16+SVMModel type=Single model2019.08 | 89 | 99 | 88.8 | 95.9 | 93.8 | 73.1 | 92.1 | 85.1 | 97.8 | 79.5 | 91.1 | 83.3 | 97.2 | 96.3 | 94.5 | 96.9 | 63.1 | 93.4 | 75 | 97.1 | 87.1 | — | |
| VGG19+SVMModel type=Single model2019.08 | 89 | 99.1 | 88.7 | 95.7 | 93.9 | 73.1 | 92.1 | 84.8 | 97.7 | 79.1 | 90.7 | 83.2 | 97.3 | 96.2 | 94.3 | 96.9 | 63.4 | 93.2 | 74.6 | 97.3 | 87.9 | — | |
| Very Deep2020.12 | 89 | 99.1 | 88.7 | 95.7 | 93.9 | 73.1 | 92.1 | 84.8 | 97.7 | 79.1 | 90.7 | 83.2 | 97.3 | 96.2 | 94.3 | 96.9 | 63.4 | 93.2 | 74.6 | 97.3 | 87.9 | — | |
| Very Deep2020.12 | 89 | 99.1 | 88.7 | 95.7 | 93.9 | 73.1 | 92.1 | 84.8 | 97.7 | 79.1 | 90.7 | 83.2 | 97.3 | 96.2 | 94.3 | 96.9 | 63.4 | 93.2 | 74.6 | 97.3 | 87.9 | — | |
| BoostLUSetup=LargeLoss [44]2025.07 | 88.69 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SPLCSetup=SPLC [9]2025.07 | 88.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| HillSetup=SPLC [9]2025.07 | 87.98 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BoostLUSetup=SPLC [9]2025.07 | 87.77 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LargeLossSetup=SPLC [9]2025.07 | 87.52 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SPLCSetup=LargeLoss [44]2025.07 | 86.53 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| [19]CNN=very-deep2015.04 | 86.3 | 96.7 | 88.8 | 92 | 87.4 | 64.7 | 91.1 | 87.4 | 94.4 | 74.9 | 89.2 | 76.3 | 93.7 | 95.2 | 91.1 | 97.6 | 66.2 | 91.2 | 70 | 94.5 | 83.7 | — | |
| ROLESetup=LargeLoss [44]2025.07 | 86.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ROLESetup=SPLC [9]2025.07 | 86.19 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| FeV-LV-20Architecture=8-layer CNN, Pre-training=ILSVRC 2012, Categories=20 bounding boxes2015.04 | 86 | 97.4 | 88.9 | 91.2 | 87.4 | 64.2 | 92.2 | 86.4 | 95 | 75.1 | 84.6 | 78.7 | 93.1 | 91.9 | 93.1 | 96.6 | 67.3 | 86.2 | 69.4 | 95.3 | 85.8 | — | |
| HillSetup=LargeLoss [44]2025.07 | 85.75 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| FeV-LV-10Architecture=8-layer CNN, Pre-training=ILSVRC 2012, Categories=10 bounding boxes2015.04 | 85.7 | 97.3 | 89.1 | 91.5 | 88.5 | 66.7 | 92.2 | 87.2 | 94 | 74 | 82.7 | 77.8 | 91.6 | 91.1 | 92.7 | 95.7 | 66.5 | 85.4 | 69.4 | 95.6 | 85.4 | — | |
| LSANSetup=LargeLoss [44]2025.07 | 85.19 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LSANSetup=SPLC [9]2025.07 | 85.13 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LargeLossSetup=LargeLoss [44]2025.07 | 84.46 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RMICModel type=Single model2019.08 | 84.4 | 98 | 85.5 | 92.6 | 88.7 | 64 | 86.8 | 82 | 94.9 | 72.7 | 83.1 | 73.4 | 95.2 | 91.7 | 90.8 | 95.5 | 58.3 | 87.6 | 70.6 | 93.8 | 83 | — | |
| RMIC2020.12 | 84.4 | 98 | 85.5 | 92.6 | 88.7 | 64 | 86.8 | 82 | 94.9 | 72.7 | 83.1 | 73.4 | 95.2 | 91.7 | 90.8 | 95.5 | 58.3 | 87.6 | 70.6 | 93.8 | 83 | — | |
| RMIC2020.12 | 84.4 | 98 | 85.5 | 92.6 | 88.7 | 64 | 86.8 | 82 | 94.9 | 72.7 | 83.1 | 73.4 | 95.2 | 91.7 | 90.8 | 95.5 | 58.3 | 87.6 | 70.6 | 93.8 | 83 | — | |
| HCP-2000CAdditional training data=2000 categories2015.04 | 84.2 | 97.5 | 84.3 | 93 | 89.4 | 62.5 | 90.2 | 84.6 | 94.8 | 69.7 | 90.2 | 74.1 | 93.4 | 93.7 | 88.8 | 93.3 | 59.7 | 90.3 | 61.8 | 94.4 | 78 | — | |
| FeVArchitecture=8-layer CNN, Pre-training=ILSVRC 20122015.04 | 84 | 96.8 | 87.8 | 88.7 | 87.2 | 63.8 | 92.3 | 86.2 | 92.3 | 72.4 | 82 | 76 | 91.9 | 90.3 | 90.3 | 95.2 | 61.2 | 82.6 | 65.6 | 92.8 | 84.4 | — | |
| PRE-1512CAdditional training data=1512 categories2015.04 | 82.8 | 94.6 | 82.9 | 88.2 | 84.1 | 60.3 | 89 | 84.4 | 90.7 | 72.1 | 86.8 | 69 | 92.1 | 93.4 | 88.6 | 96.1 | 64.3 | 86.6 | 62.3 | 91.1 | 79.8 | — | |
| NUS-PSLArchitecture=hand-crafted feature based2015.04 | 82.2 | 97.3 | 84.2 | 80.8 | 85.3 | 60.8 | 89.9 | 86.8 | 89.3 | 75.4 | 77.8 | 75.1 | 83 | 87.5 | 90.1 | 87.6 | 59 | 79.2 | 73.4 | 94.5 | 80.7 | — | |
| HCP-1000CArchitecture=8-layer CNN, Pre-training=ILSVRC 20122015.04 | 81.7 | 97.7 | 83 | 93.2 | 87.2 | 59.6 | 88.2 | 81.9 | 94.7 | 66.9 | 81.6 | 68 | 93 | 88.2 | 92.7 | 95 | 57.8 | 85.1 | 55.4 | 93 | 77.2 | — | |
| PRE-1000CArchitecture=8-layer CNN, Pre-training=ILSVRC 20122015.04 | 78.7 | 93.5 | 78.4 | 87.7 | 80.9 | 57.3 | 85 | 81.6 | 89.4 | 66.9 | 73.8 | 62 | 89.5 | 83.2 | 87.7 | 95.8 | 61.4 | 79 | 54.3 | 88 | 78.3 | — | |
| CLIP2026.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 74.25 | |
| ML-TTA2026.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 76.63 | |
| SSA2026.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 77.29 | |
| TPT2026.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 73.92 |