Image-to-Text matching on MSCOCO
78.4R@1GSMN (sparse+dense)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| GSMN (sparse+dense)Image Backbone=Faster R-CNN, Text Backbone=ft Bi-GRU2020.04 | 78.4 | 96.4 | 98.6 | 522.5 | |
| PFANImage Backbone=Faster R-CNN, Text Backbone=ft Bi-GRU2020.04 | 76.5 | 96.3 | 99 | 518.2 | |
| GSMN (sparse)Image Backbone=Faster R-CNN, Text Backbone=ft Bi-GRU2020.04 | 76.1 | 95.6 | 98.3 | 514 | |
| BFANImage Backbone=Faster R-CNN, Text Backbone=ft Bi-GRU2020.04 | 74.9 | 95.2 | — | 317.9 | |
| GSMN (dense)Image Backbone=Faster R-CNN, Text Backbone=ft Bi-GRU2020.04 | 74.7 | 95.3 | 98.2 | 511.6 | |
| SCANImage Backbone=Faster R-CNN, Text Backbone=ft Bi-GRU2020.04 | 72.7 | 94.8 | 98.4 | 507.9 | |
| SCOImage Backbone=fixed ResNet-152, Text Backbone=ft LSTM2020.04 | 69.9 | 92.9 | 97.5 | 499.3 | |
| GXNImage Backbone=ft ResNet-152, Text Backbone=ft GRU2020.04 | 68.5 | — | 97.9 | 317.5 | |
| DPCImage Backbone=ft ResNet-152, Text Backbone=ft ResNet-1522020.04 | 65.5 | 89.8 | 95.5 | 467.8 | |
| VSE++Image Backbone=ft ResNet-152, Text Backbone=ft GRU2020.04 | 64.7 | — | 95.9 | 304.6 | |
| DSPEImage Backbone=fixed VGG-19, Text Backbone=w2v+HGLMM2020.04 | 50.1 | 79.7 | 89.2 | 420.7 | |
| m-CNNImage Backbone=fixed VGG-19, Text Backbone=ft CNN2020.04 | 42.8 | 73.1 | 84.1 | 384 |