Cross-modal Retrieval on MSCOCO 5K (test)
0.621R@1 (I2T)DSMD
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| DSMDBackbone=ViT, BERT, Visual Embed=Image, Ensemble=false2024.04 | 0.621 | 0.859 | 0.92 | 0.48 | 0.756 | 0.845 | 448.1 | |
| RCARBackbone=BUTD, GRU, Visual Embed=Region, Ensemble=true2024.04 | 0.613 | 0.861 | 0.926 | 0.443 | 0.732 | 0.832 | 440.7 | |
| DIMEBackbone=BUTD, BERT, Visual Embed=Region, Ensemble=true2024.04 | 0.593 | 0.854 | 0.919 | 0.431 | 0.73 | 0.831 | 435.8 | |
| GPOBackbone=BUTD, BERT, Visual Embed=Region, Ensemble=false2024.04 | 0.583 | 0.853 | 0.923 | 0.424 | 0.727 | 0.832 | 434.2 | |
| SGRAFBackbone=BUTD, GRU, Visual Embed=Region, Ensemble=false2024.04 | 0.578 | — | 0.916 | 0.419 | — | 0.813 | — | |
| CAMERABackbone=BUTD, BERT, Visual Embed=Region, Ensemble=true2024.04 | 0.551 | 0.829 | 0.912 | 0.405 | 0.717 | 0.825 | 423.9 | |
| VSRN++Backbone=BUTD, BERT, Visual Embed=Region, Ensemble=false2024.04 | 0.547 | 0.829 | 0.909 | 0.42 | 0.722 | 0.827 | 425.4 | |
| MMCABackbone=BUTD, BERT, Visual Embed=Region, Ensemble=false2024.04 | 0.54 | 0.825 | 0.907 | 0.387 | 0.697 | 0.808 | 416.4 | |
| IMRAMBackbone=BUTD, GRU, Visual Embed=Region, Ensemble=false2024.04 | 0.537 | 0.832 | 0.91 | 0.396 | 0.691 | 0.798 | 416.4 | |
| VSRNBackbone=BUTD, GRU, Visual Embed=Region, Ensemble=true2024.04 | 0.53 | 0.811 | 0.894 | 0.405 | 0.706 | 0.811 | 415.7 | |
| CAANBackbone=BUTD, GRU, Visual Embed=Region, Ensemble=false2024.04 | 0.525 | 0.833 | 0.909 | 0.412 | 0.703 | 0.829 | 421.1 | |
| SCANBackbone=BUTD, GRU, Visual Embed=Region, Ensemble=true2024.04 | 0.504 | 0.822 | 0.9 | 0.386 | 0.693 | 0.804 | 410.9 |