Image-to-Text Retrieval on COCO 5K Full (test)
71R@1Ours†
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Ours†Backbone=ResNeXt-101 + BERT, CA (Cross-Attention)=false, Ensemble=true2022.11 | 71 | 91.8 | 96.3 | 482 | |
| OursBackbone=ResNeXt-101 + BERT, CA (Cross-Attention)=false, Ensemble=false2022.11 | 69.1 | 90.7 | 95.6 | 474.9 | |
| VSE∞†Backbone=ResNeXt-101 + BERT, CA (Cross-Attention)=false, Ensemble=true2022.11 | 68.1 | 90.2 | 95.2 | 474.8 | |
| VSE∞Backbone=ResNeXt-101 + BERT, CA (Cross-Attention)=false, Ensemble=false2022.11 | 66.4 | 89.3 | 94.6 | 468.9 | |
| Ours†Backbone=Faster R-CNN + Bi-GRU, CA (Cross-Attention)=false, Ensemble=true2022.11 | 60.4 | 86.2 | 92.4 | 437.8 | |
| NAAFBackbone=Faster R-CNN + Bi-GRU, CA (Cross-Attention)=true, Ensemble=true2022.11 | 58.9 | 85.2 | 92 | 430.9 | |
| OursBackbone=Faster R-CNN + Bi-GRU, CA (Cross-Attention)=false, Ensemble=false2022.11 | 58.8 | 84.9 | 91.5 | 430.7 | |
| SGRAFBackbone=Faster R-CNN + Bi-GRU, CA (Cross-Attention)=true, Ensemble=true2022.11 | 57.8 | — | 91.6 | 413 | |
| VSE∞Backbone=Faster R-CNN + Bi-GRU, CA (Cross-Attention)=false, Ensemble=false2022.11 | 56.6 | 83.6 | 91.4 | 421.9 | |
| IMRAMBackbone=Faster R-CNN + Bi-GRU, CA (Cross-Attention)=true, Ensemble=true2022.11 | 53.7 | 83.2 | 91 | 416.5 | |
| VSRNBackbone=Faster R-CNN + Bi-GRU, CA (Cross-Attention)=false, Ensemble=true2022.11 | 53 | 81.1 | 89.4 | 415.7 | |
| CAANBackbone=Faster R-CNN + Bi-GRU, CA (Cross-Attention)=true, Ensemble=false2022.11 | 52.5 | 83.3 | 90.9 | 421.1 | |
| SCANBackbone=Faster R-CNN + Bi-GRU, CA (Cross-Attention)=true, Ensemble=true2022.11 | 50.4 | 82.2 | 90 | 410.9 | |
| OursBackbone=ResNet-152 + Bi-GRU, CA (Cross-Attention)=false, Ensemble=false2022.11 | 47.2 | 74.8 | 84.1 | 377.7 | |
| PVSEBackbone=ResNet-152 + Bi-GRU, CA (Cross-Attention)=false, Ensemble=false2022.11 | 45.2 | 74.3 | 84.5 | 374.4 | |
| PCMEBackbone=ResNet-152 + Bi-GRU, CA (Cross-Attention)=false, Ensemble=false2022.11 | 44.2 | — | — | — | |
| VSE++Backbone=ResNet-152 + Bi-GRU, CA (Cross-Attention)=false, Ensemble=false2022.11 | 41.3 | 71.1 | 81.2 | 355.7 |