Caption Retrieval on Flickr30k (test)
76.4R@1GSMN
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GSMN2021.06 | 76.4 | 94.3 | 97.3 | — | 268 | |
| MMCA2021.06 | 74.2 | 92.8 | 96.4 | — | 263.4 | |
| IMRAM2021.06 | 74.1 | 93 | 96.6 | — | 263.7 | |
| SGM2021.06 | 71.8 | 91.7 | 95.5 | — | 259 | |
| VSRNBackbone=Faster R-CNN, ResNet2019.09 | 71.3 | 90.6 | 96 | — | — | |
| LGSGM2021.06 | 71 | 91.9 | 96.1 | — | 259 | |
| PFAN2021.06 | 70 | 91.8 | 95 | — | 256.8 | |
| SCAN i-t AVGBackbone=Faster R-CNN, ResNet, lambda_1=4, Pooling=Average, Attention Direction=Image-Text2018.03 | 67.9 | 89 | 94.4 | — | — | |
| SCAN i-t LSEBackbone=Faster R-CNN, ResNet, lambda_1=4, lambda_2=5, Pooling=LogSumExp, Attention Direction=Image-Text2018.03 | 67.7 | 88.9 | 94 | — | — | |
| SCAN t-i AVG + i-t LSEBackbone=Faster R-CNN, ResNet, ensemble=t-i AVG and i-t LSE2018.03 | 67.4 | 90.3 | 95.8 | — | — | |
| SCANBackbone=Faster R-CNN, ResNet2019.09 | 67.4 | 90.3 | 95.8 | — | — | |
| SCAN t-i AVGBackbone=Faster R-CNN, ResNet, lambda_1=9, Pooling=Average, Attention Direction=Text-Image2018.03 | 61.8 | 87.5 | 93.7 | — | — | |
| SCAN t-i LSEBackbone=Faster R-CNN, ResNet, lambda_1=9, lambda_2=6, Pooling=LogSumExp, Attention Direction=Text-Image2018.03 | 61.1 | 85.4 | 91.5 | — | — | |
| DPCBackbone=ResNet2018.03 | 55.6 | 81.9 | 89.5 | — | — | |
| SCOBackbone=ResNet2018.03 | 55.5 | 82 | 89.3 | — | — | |
| SCOBackbone=ResNet2019.09 | 55.5 | 82 | 89.3 | — | — | |
| DANBackbone=ResNet2018.03 | 55 | 81.8 | 89 | — | — | |
| VSE++Backbone=ResNet2018.03 | 52.9 | — | 87.2 | — | — | |
| VSE++Backbone=ResNet2019.09 | 52.9 | 79.1 | 87.2 | — | — | |
| 2WayNetBackbone=VGG2018.03 | 49.8 | 67.5 | — | — | — | |
| 2WayNetBackbone=VGG2019.09 | 49.8 | 67.5 | — | — | — | |
| RRFBackbone=ResNet2019.09 | 47.6 | 77.4 | 87.1 | — | — | |
| SM-LSTMBackbone=VGG2018.03 | 42.5 | 71.9 | 81.5 | — | — | |
| SMlstmBackbone=VGG2019.09 | 42.5 | 71.9 | 81.5 | — | — | |
| HM-LSTMBackbone=R-CNN, AlexNet2018.03 | 38.1 | — | 76.5 | — | — | |
| HMIstmBackbone=R-CNN, AlexNet2019.09 | 38.1 | — | 76.5 | — | — | |
| FVBackbone=VGG2019.09 | 35 | 62 | 73.8 | — | — | |
| VQABackbone=VGG2019.09 | 33.9 | 62.5 | 74.5 | — | — | |
| LRCNconfiguration=factored, visual_model=CaffeNet2014.11 | 23.6 | 46.6 | 58.3 | 7 | — | |
| DVSABackbone=R-CNN, AlexNet2018.03 | 22.2 | 48.2 | 61.4 | — | — | |
| DVSABackbone=R-CNN, AlexNet2019.09 | 22.2 | 48.2 | 61.4 | — | — | |
| m-RNN2014.11 | 18.4 | 40.2 | 50.9 | 10 | — | |
| DeFrag2014.11 | 16.4 | 40.2 | 54.7 | 8 | — | |
| ConvNetarchitecture=AlexNet2014.11 | 14.8 | 39.2 | 50.9 | 10 | — | |
| SDT-RNN2014.11 | 9.6 | 29.8 | 41.1 | 16 | — | |
| DeViSE2014.11 | 4.5 | 18.1 | 29.2 | 26 | — |