Image-to-Text Retrieval on CUHK-PEDES (test)
84.92Rank-1VGSG
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| VGSGArch.=ViT-Base, Reference=Ours2023.11 | 84.92 | 96.35 | 98.24 | 63.83 | — | |
| VGSGArch.=ResNet50, Reference=Ours2023.11 | 82.49 | 95.64 | 97.78 | 61.42 | — | |
| CM-MoCo (ResNet101)Arch=Single-scale, Dim=256, Reranking=false2021.10 | 78.99 | 95.02 | 97.17 | 56.78 | — | |
| CM-MoCo (ResNet101)Arch=Single-scale, Dim=256, Reranking=true2021.10 | 78.99 | 95.02 | 97.23 | 57.31 | — | |
| CM-MoCo (ResNet50)Arch=Single-scale, Dim=256, Reranking=true2021.10 | 76.26 | 93.46 | 96.58 | 55.67 | — | |
| CM-MoCo (ResNet50)Arch=Single-scale, Dim=256, Reranking=false2021.10 | 75.96 | 93.4 | 96.55 | 55.05 | — | |
| TextReIDArch.=ResNet50, Reference=BMVC'212023.11 | 75.96 | 93.4 | 96.55 | 55.05 | — | |
| TIPCBArch=Multi-scale, Dim=2048, Reranking=true2021.10 | 74.04 | 92.06 | 95.61 | 53.78 | — | |
| TIPCBArch=Multi-scale, Dim=2048, Reranking=false2021.10 | 73.55 | 92.26 | 96.03 | 51.78 | — | |
| TIPCBArch.=ResNet50, Reference=Neuro'222023.11 | 73.55 | 92.26 | 96.03 | 51.78 | — | |
| NAFS (G)Arch=Multi-scale, Dim=768, Reranking=true2021.10 | 72.67 | 90.92 | 95.12 | 50.92 | — | |
| NAFS (G)Arch=Multi-scale, Dim=768, Reranking=false2021.10 | 71.89 | 90.99 | 95.28 | 50.16 | — | |
| NAFS (G)Arch.=ResNet50, Reference=arXiv'212023.11 | 71.89 | 90.99 | 95.28 | 50.16 | — | |
| MGELArch=Multi-scale, Dim=512, Reranking=false2021.10 | 71.87 | 91.38 | 95.42 | — | — | |
| MGELArch.=ResNet50, Reference=IJCAI'212023.11 | 71.87 | 91.38 | 95.42 | — | — | |
| HGANArch=Multi-scale, Dim=512, Reranking=false2021.10 | 71.16 | 90.05 | 95.06 | — | — | |
| HGANArch.=ResNet50, Reference=MM'202023.11 | 71.16 | 90.05 | 95.06 | — | — | |
| TIMAM+Arch=Single-scale, Dim=512, Reranking=false2021.10 | 67.4 | 88.65 | 93.91 | — | — | |
| TIMAMArch.=ResNet101, Reference=ICCV'192023.11 | 67.4 | 88.65 | 93.91 | — | — | |
| ViTAAArch=Multi-scale, Dim=256, Reranking=true2021.10 | 66.17 | 88.61 | 93.56 | 46.39 | — | |
| ViTAAArch=Multi-scale, Dim=256, Reranking=false2021.10 | 65.71 | 88.68 | 93.75 | 45.75 | — | |
| ViTAAArch.=ResNet50, Reference=ECCV'202023.11 | 65.71 | 88.68 | 93.75 | 45.75 | — | |
| CMPM/C+Arch=Single-scale, Dim=512, Reranking=false2021.10 | 60.96 | 84.42 | 90.83 | — | — | |
| CMPM/CArch.=MobileNet, Reference=ECCV'20182023.11 | 60.96 | 84.42 | 90.83 | — | — | |
| Baseepsilon=12026.04 | — | — | — | — | 15.5 | |
| Baseepsilon=102026.04 | — | — | — | — | 15.5 | |
| DP-CLIPepsilon=12026.04 | — | — | — | — | 15.1 | |
| DP-CLIPepsilon=102026.04 | — | — | — | — | 18.5 | |
| DP-GCLepsilon=12026.04 | — | — | — | — | 55.8 | |
| DP-GCLepsilon=102026.04 | — | — | — | — | 65.9 | |
| DP-SGDepsilon=12026.04 | — | — | — | — | 14.5 | |
| DP-SGDepsilon=102026.04 | — | — | — | — | 14.2 | |
| Logit-DPepsilon=12026.04 | — | — | — | — | 16.1 | |
| Logit-DPepsilon=102026.04 | — | — | — | — | 16.4 |