Multi-label recognition on MS-COCO (val)
81.5F1 Score (All)PLA
Evaluation results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| PLAtraining resolution=288, testing resolution=2882021.10 | 81.5 | 80.4 | 68.9 | 74.2 | 73.3 | 77.2 | — | — | — | — | — | — | — | |
| Transformer-based Dual Relation Graphtraining resolution=576, testing resolution=5762021.10 | 80.4 | 86 | 87 | 74.7 | 87.5 | 77.9 | 82.4 | 90.7 | 65.6 | 76.2 | 91.9 | 68 | 78.1 | |
| ADD-GCNtraining resolution=448, testing resolution=5762021.10 | 80.1 | 85.2 | 84.7 | 75.9 | 84.9 | 79.4 | 82 | 88.8 | 66.2 | 75.8 | 90.3 | 68.5 | 77.9 | |
| Transformer-based Dual Relation Graphtraining resolution=448, testing resolution=5762021.10 | 80.1 | 85.8 | 87.9 | 73.6 | 87.9 | 77.3 | 82.3 | 91.3 | 64.8 | 75.8 | 92 | 67.6 | 77.9 | |
| C-Trantraining resolution=576, testing resolution=5762021.10 | 79.9 | 85.1 | 86.3 | 74.3 | 87.7 | 76.5 | 81.7 | 90.1 | 65.7 | 76 | 92.1 | 71.4 | 77.6 | |
| Transformer-based Dual Relation Graphtraining resolution=448, testing resolution=4482021.10 | 79 | 84.6 | 86 | 73.1 | 86.6 | 76.4 | 81.2 | 89.9 | 64.4 | 75 | 91.2 | 67 | 77.2 | |
| ML-GCNtraining resolution=448, testing resolution=4482021.10 | 78 | 83 | 85.1 | 72 | 85.8 | 75.4 | 80.3 | 89.2 | 64.1 | 74.6 | 90.5 | 66.5 | 76.7 | |
| KSSNettraining resolution=448, testing resolution=4482021.10 | 77.2 | 83.7 | 84.6 | 73.2 | 87.8 | 76.2 | 81.5 | — | — | — | — | — | — | |
| SSGRLtraining resolution=576, testing resolution=5762021.10 | 76.8 | 83.8 | 89.9 | 68.5 | 91.3 | 70.8 | 79.7 | 91.9 | 62.5 | 72.7 | 93.8 | 64.1 | 76.2 | |
| ResNet-101 (ours implementation)training resolution=448, testing resolution=4482021.10 | 73 | 78.6 | 82.4 | 65.5 | 86 | 70.4 | 77.4 | 85.9 | 58.6 | 69.7 | 90.5 | 62.8 | 74.1 | |
| DSRBBackbone=ResNet-101, Resolution=448x4482022.05 | 72.7 | — | — | — | — | — | 76.8 | — | — | — | — | — | — | |
| SRNtraining resolution=224, testing resolution=2242021.10 | 71.2 | 77.1 | 81.6 | 65.4 | 82.7 | 69.9 | 75.8 | 85.2 | 58.8 | 67.4 | 87.4 | 62.5 | 72.9 | |
| KGGRBackbone=ResNet-101, Resolution=448x4482022.05 | 69.7 | — | — | — | — | — | 73.7 | — | — | — | — | — | — | |
| Partial BCEBackbone=ResNet-101, Resolution=448x4482022.05 | 68.8 | — | — | — | — | — | 74 | — | — | — | — | — | — | |
| GCN-MLBackbone=ResNet-101, Resolution=448x4482022.05 | 68.4 | — | — | — | — | — | 73.1 | — | — | — | — | — | — | |
| SSGRLBackbone=ResNet-101, Resolution=448x4482022.05 | 68.1 | — | — | — | — | — | 73.9 | — | — | — | — | — | — | |
| CLBackbone=ResNet-101, Resolution=448x4482022.05 | 48.3 | — | — | — | — | — | 61.9 | — | — | — | — | — | — | |
| ASLBackbone=ResNet-101, Resolution=448x4482022.05 | 47.9 | — | — | — | — | — | 46.7 | — | — | — | — | — | — | |
| CNN-RNN2021.10 | — | 61.2 | — | — | — | — | — | 66 | 55.6 | 60.4 | 69.2 | 66.4 | 67.8 | |
| Order-Free RNN2021.10 | — | — | — | — | — | — | — | 71.6 | 54.8 | 62.1 | 74.2 | 62.2 | 67.7 | |
| RNN-Attention2021.10 | — | — | — | — | — | — | — | 79.1 | 58.7 | 67.4 | 84 | 63 | 72 | |
| TaI promptingDPT=false, Zero-shot=true2022.11 | — | 61.1 | — | — | — | — | — | — | — | — | — | — | — | |
| TaI-DPTDPT=true, Zero-shot=true2022.11 | — | 65.1 | — | — | — | — | — | — | — | — | — | — | — | |
| ZSCLIPDPT=false, Zero-shot=true2022.11 | — | 47.3 | — | — | — | — | — | — | — | — | — | — | — | |
| ZSCLIPDPT=true, Zero-shot=true2022.11 | — | 49.7 | — | — | — | — | — | — | — | — | — | — | — |