Image Annotation on Flickr30k (test)
55.5R@1semantic-enhanced image and sentence matching model
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| semantic-enhanced image and sentence matching modelBackbone=ResNet-1522017.12 | 55.5 | 89.3 | 69.7 | 82 | — | |
| DANBackbone=ResNet-1522017.12 | 55 | 89 | 68.9 | 81.8 | — | |
| VSE++Backbone=ResNet-1522017.12 | 52.9 | 87.2 | 68 | 79.1 | — | |
| 2WayNetBackbone=VGGNet-192017.12 | 49.8 | — | — | 67.5 | — | |
| RRFBackbone=ResNet-1522017.12 | 47.6 | 87.1 | 66 | 77.4 | — | |
| semantic-enhanced image and sentence matching modelBackbone=VGGNet-192017.12 | 44.2 | 83.6 | 62.3 | 74.1 | — | |
| sm-LSTMensemble=true2016.11 | 42.5 | 81.5 | 2 | 71.9 | 358.7 | |
| sm-LSTMBackbone=VGGNet-192017.12 | 42.5 | 81.5 | 59.8 | 71.9 | — | |
| sm-LSTM2016.11 | 42.4 | 79.9 | 2 | 67.5 | 343.4 | |
| DANBackbone=VGGNet-192017.12 | 41.4 | 82.5 | 60.6 | 73.5 | — | |
| VSE++Backbone=VGGNet-192017.12 | 41.3 | 77.9 | 58.4 | 69 | — | |
| DSPE+FVuses external text corpora=true2016.11 | 40.3 | 79.9 | — | 68.9 | 351 | |
| DSPEBackbone=VGGNet-192017.12 | 40.3 | 79.9 | 58.5 | 68.9 | — | |
| RTP+FVensemble=true, uses external text corpora=true2016.11 | 37.4 | 74.3 | — | 63.1 | 326.1 | |
| RTPBackbone=VGGNet-192017.12 | 37.4 | 74.3 | 54.3 | 63.1 | — | |
| m-RNN2016.11 | 35.4 | 73.7 | 3 | 63.8 | 309.5 | |
| m-RNNBackbone=VGGNet-192017.12 | 35.4 | 73.7 | 51.6 | 63.8 | — | |
| FVensemble=true, uses external text corpora=true2016.11 | 35 | 73.8 | 3 | 62 | 314.5 | |
| FVBackbone=VGGNet-192017.12 | 35 | 73.8 | 52.4 | 62 | — | |
| RNN+FVuses external text corpora=true2016.11 | 34.7 | 72.6 | 3 | 62.7 | 320.5 | |
| RNN+FVBackbone=VGGNet-192017.12 | 34.7 | 72.6 | 53.4 | 62.7 | — | |
| VQABackbone=VGGNet-192017.12 | 33.9 | 74.5 | 52.2 | 62.5 | — | |
| m-CNNensemble=true2016.11 | 33.6 | 74.9 | 3 | 64.1 | 324.7 | |
| m-CNNBackbone=VGGNet-192017.12 | 33.6 | 74.9 | 54.1 | 64.1 | — | |
| sm-LSTM-ctx2016.11 | 33.5 | 70.8 | 3 | 60.6 | 300.1 | |
| sm-LSTM-att2016.11 | 27 | 65.6 | 5 | 53.6 | 271.1 | |
| sm-LSTM-mean2016.11 | 25.9 | 65.4 | 5 | 53.1 | 261.5 | |
| MNLM2014.11 | 23 | 63 | 5 | — | — | |
| MNLM2016.11 | 23 | 62.9 | 5 | 50.7 | 251.9 | |
| MNLMBackbone=VGGNet-192017.12 | 23 | 62.9 | 42 | 50.7 | — | |
| DVSA (BRNN)2016.11 | 22.2 | 61.4 | 4.8 | 48.2 | 235.2 | |
| DVSABackbone=VGGNet-192017.12 | 22.2 | 61.4 | 39.2 | 48.2 | — | |
| m-RNN2014.11 | 18 | 51 | 10 | — | — | |
| NIC2014.11 | 17 | 56 | 7 | — | — | |
| NIC2016.11 | 17 | 56 | 7 | — | — | |
| DCCA2016.11 | 16.7 | 52.9 | 8 | 39.3 | 195.5 | |
| DeFrag2014.11 | 16 | 55 | 8 | — | — | |
| Deep Fragment2016.11 | 14.2 | 51.3 | 10 | 37.7 | 188.4 | |
| RVP(T+I)2016.11 | 12.1 | 47.8 | 11 | 27.8 | 178.4 |