Referring Expression Comprehension on ReferItGame (test)
79.3Top-1 AccSimVG-TB
Evaluation Results
| Method | Links | |
|---|---|---|
| SimVG-TBVisual Encoder=ViT-L/32, Time (ms)=1012024.09 | 79.3 | |
| SimVG-DBVisual Encoder=ViT-L/32, Time (ms)=1162024.09 | 78.75 | |
| Referring TransformerBackbone=RN101, Stage=One-stage, Inference time (ms)=40(15), Pre-trained=true2021.06 | 76.18 | |
| Referring TransformerBackbone=RN50, Stage=One-stage, Inference time (ms)=37(14), Pre-trained=true2021.06 | 75.49 | |
| SimVG-DBVisual Encoder=ViT-B/32, Time (ms)=522024.09 | 74.83 | |
| QRNetBackbone=Swin-S/BERT2024.06 | 74.61 | |
| SimVG-TBVisual Encoder=ViT-B/32, Time (ms)=442024.09 | 74.59 | |
| TransCPVisual Encoder=RN50, Time (ms)=74*2024.09 | 72.05 | |
| VLTVGZero-Shot=false2024.03 | 71.98 | |
| PLV-FPNBackbone=ResNet-101/BERT2024.06 | 71.77 | |
| VLTVGVisual Encoder=RN50, Time (ms)=79*2024.09 | 71.6 | |
| Referring TransformerBackbone=RN101, Stage=One-stage, Inference time (ms)=40(15), Pre-trained=false2021.06 | 71.42 | |
| RefTRBackbone=ResNet-101/BERT2024.06 | 71.42 | |
| LADSBackbone=ResNet-50/BERT2024.06 | 71.08 | |
| CLIP-VGZero-Shot=false2024.03 | 70.89 | |
| Referring TransformerBackbone=RN50, Stage=One-stage, Inference time (ms)=37(14), Pre-trained=false2021.06 | 70.81 | |
| TransVGBackbone=ResNet-101/BERT2024.06 | 70.73 | |
| TransVGVisual Encoder=RN101, Time (ms)=622024.09 | 70.73 | |
| Dyn.MDETRVisual Encoder=ViT-B/162024.09 | 70.37 | |
| SeqTRZero-Shot=false2024.03 | 69.66 | |
| SeqTRBackbone=DarkNet-53/GRU2024.06 | 69.66 | |
| SeqTRVisual Encoder=DN53, Time (ms)=502024.09 | 69.66 | |
| Qwen-VL + SC-TuneZero-Shot=true2024.03 | 69.28 | |
| ScanFormerBackbone=Unified Transformer2024.06 | 68.85 | |
| PFOSBackbone=ResNet-101/BERT2024.06 | 67.9 | |
| LBYLVisual Encoder=DN53, Time (ms)=302024.09 | 67.47 | |
| NMTreeVisual Encoder=RN1012024.09 | 66.44 | |
| ReSC-LargeBackbone=DN53, Stage=One-stage, Inference time (ms)=362021.06 | 64.6 | |
| ReSC-LargeBackbone=DarkNet-53/BERT2024.06 | 64.6 | |
| ReSCLVisual Encoder=DN53, Time (ms)=362024.09 | 64.6 | |
| RCCFBackbone=DLA34, Stage=One-stage, Inference time (ms)=252021.06 | 63.79 | |
| VGTRVisual Encoder=RN502024.09 | 63.63 | |
| DDPNBackbone=RN101, Stage=Two-stage2021.06 | 63 | |
| Qwen-VLZero-Shot=true2024.03 | 61.48 | |
| FAOABackbone=DN53, Stage=One-stage, Inference time (ms)=232021.06 | 60.67 | |
| FAOABackbone=DarkNet-53/BERT2024.06 | 60.67 | |
| FAOAVisual Encoder=DN53, Time (ms)=392024.09 | 60.67 | |
| ZSGNetBackbone=RN50, Stage=One-stage2021.06 | 58.63 | |
| SSGBackbone=DN53, Stage=One-stage, Inference time (ms)=252021.06 | 54.24 | |
| DGAVisual Encoder=VGG16, Time (ms)=3412024.09 | 51.99 | |
| MiniGPT-v2 + SC-TuneZero-Shot=true2024.03 | 43.68 | |
| MiniGPT-v2Zero-Shot=true2024.03 | 36.05 | |
| CITEBackbone=RN101, Stage=Two-stage, Inference time (ms)=1962021.06 | 35.07 | |
| Similarity NetBackbone=RN101, Stage=Two-stage, Inference time (ms)=1842021.06 | 34.54 | |
| MAttNetBackbone=RN101, Stage=Two-stage, Inference time (ms)=3202021.06 | 29.04 | |
| MAttNetBackbone=ResNet-101/LSTM2024.06 | 29.04 | |
| MAttNetVisual Encoder=RN101, Time (ms)=3202024.09 | 29.04 |