Composed Image Retrieval on Shoes (test)
70.81R@10COMBINER
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| COMBINERModel Category=VLP Model-Based Methods2026.06 | 70.81 | 32.37 | 88.81 | 64 | |
| DQU-CIRModel Category=VLP Model-Based Methods2026.06 | 69.19 | 31.47 | 88.52 | 63.06 | |
| LIMN+Model Category=VLP Model-Based Methods2026.06 | 68.37 | — | 88.07 | — | |
| LIMNModel Category=VLP Model-Based Methods2026.06 | 68.2 | — | 87.45 | — | |
| ENCODERModel Category=VLP Model-Based Methods2026.06 | 65.59 | 26.97 | 86.48 | 59.68 | |
| TG-CIRModel Category=VLP Model-Based Methods2026.06 | 63.2 | 25.89 | 85.07 | 58.05 | |
| Prog. Lrn.Model Category=VLP Model-Based Methods2026.06 | 58.83 | 22.88 | 84.16 | 55.29 | |
| IUDCModel Category=VLP Model-Based Methods2026.06 | 56.82 | 21.17 | 82.25 | 53.41 | |
| Css-Net2023.06 | 56.81 | 20.13 | 81.32 | — | |
| EERModel Category=Traditional Model-Based Methods2026.06 | 56.02 | 20.05 | 79.94 | 52 | |
| CRNModel Category=Traditional Model-Based Methods2026.06 | 54.55 | 18.92 | 80.04 | 51.17 | |
| CLVC-Netensemble=true2023.06 | 54.39 | 17.64 | 79.47 | — | |
| DCNetVisual Backbone=ResNet-502022.11 | 53.82 | — | 79.33 | — | |
| DCNet2023.06 | 53.82 | — | 79.33 | — | |
| Multi-grained Uncertainty RegularizationVisual Backbone=ResNet-502022.11 | 53.63 | 18.41 | 79.84 | — | |
| MUR2023.06 | 53.63 | 18.41 | 79.84 | — | |
| MGURModel Category=Traditional Model-Based Methods2026.06 | 53.63 | 18.41 | 79.84 | 50.63 | |
| ARTEMISImage Backbone=ResNet50, Text Backbone=BiGRU2022.03 | 53.11 | 18.72 | 79.31 | 50.38 | |
| ARTEMISVisual Backbone=ResNet-502022.11 | 53.11 | 18.72 | 79.31 | — | |
| ARTEMIS2023.06 | 53.11 | 18.72 | 79.31 | — | |
| Baseline (CoSMo + RoBERTa)text encoder=RoBERTa2023.06 | 52.26 | 17.27 | 77.35 | — | |
| VALImage Backbone=ResNet50, Additional Information=true2022.03 | 51.52 | 17.18 | 75.83 | 48.18 | |
| VALModel Category=Traditional Model-Based Methods2026.06 | 51.52 | 17.18 | 75.83 | 48.18 | |
| ARTEMISImage Backbone=ResNet50, Text Backbone=LSTM2022.03 | 51.05 | 17.6 | 76.85 | 48.5 | |
| BaselineVisual Backbone=ResNet-502022.11 | 49.48 | 15.26 | 76.46 | — | |
| VALImage Backbone=ResNet502022.03 | 49.12 | 16.49 | 73.53 | 46.38 | |
| VALVisual Backbone=ResNet-502022.11 | 49.12 | 16.49 | 73.53 | — | |
| VAL2023.06 | 49.12 | 16.49 | 73.53 | — | |
| TIRGImage Backbone=ResNet50, Text Backbone=LSTM, Re-implementation=true2022.03 | 48.65 | 15.52 | 76.49 | 46.89 | |
| CoSMoImage Backbone=ResNet502022.03 | 48.36 | 16.72 | 75.64 | 46.91 | |
| CoSMoVisual Backbone=ResNet-502022.11 | 48.36 | 16.72 | 75.64 | — | |
| CoSMo2023.06 | 48.36 | 16.72 | 75.64 | — | |
| TIRGImage Backbone=ResNet50, Text Backbone=BiGRU, Re-implementation=true2022.03 | 47.51 | 14.46 | 75.17 | 45.71 | |
| TIRGVisual Backbone=ResNet-172022.11 | 45.45 | 12.6 | 69.39 | — | |
| TIRG2023.06 | 45.45 | 12.6 | 69.39 | — | |
| TIRGModel Category=Traditional Model-Based Methods2026.06 | 45.45 | 12.6 | 69.39 | 42.48 | |
| AFCEModel Category=VLP Model-Based Methods2026.06 | 44.31 | 19.1 | 55.37 | 79.57 | |
| MRNVisual Backbone=ResNet-1522022.11 | 41.7 | 11.74 | 67.01 | — | |
| MRN2023.06 | 41.7 | 11.74 | 67.01 | — | |
| FiLMVisual Backbone=ResNet-502022.11 | 38.89 | 10.19 | 68.3 | — | |
| FiLM2023.06 | 38.89 | 10.19 | 68.3 | — |