Referring Image Segmentation on G-Ref (val)
75.54mIoUVATEX
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| VATEXBackbone=Swin-B, Training=Joint2024.04 | 75.54 | — | — | — | — | — | |
| VATEX_jointExternal Datasets=RefCOCO(+/g) (368K), Backbone=Swin-B2024.04 | 75.54 | — | — | — | — | — | |
| HIPIEExternal Datasets=Object365 (30M) & COCO + RefCOCO(+/g)2024.04 | 75.3 | — | — | — | — | — | |
| UNINEXT-HExternal Datasets=Object365 (30M) & COCO + RefCOCO(+/g)2024.04 | 74.7 | — | — | — | — | — | |
| PaLI-3resolution=812x812, fine-tuned=true2023.10 | 72.72 | — | — | — | — | — | |
| PolyFormer2023.10 | 71.15 | — | — | — | — | — | |
| VATEXBackbone=Swin-B2024.04 | 69.73 | — | — | — | — | — | |
| VATEXVisual Backbone=Swin-B, Textual Backbone=CLIP2024.04 | 69.73 | — | — | — | — | — | |
| VATEX_G-RefExternal Datasets=G-Ref (85K), Backbone=Swin-B2024.04 | 69.73 | — | — | — | — | — | |
| PolyFormer-BExternal Datasets=RefCOCO(+/g) (368K)2024.04 | 69.36 | — | — | — | — | — | |
| SEEM2024.04 | 67.8 | — | — | — | — | — | |
| X-Decoder2024.04 | 67.5 | — | — | — | — | — | |
| RefTr2023.10 | 66.63 | — | — | — | — | — | |
| RefTRExternal Datasets=Visual Genome (5.4M) & Flickr30k-entities (158K)2024.04 | 66.63 | — | — | — | — | — | |
| LISA-7BVisual Backbone=ViT-H SAM, Textual Backbone=Vicuna-7B2024.04 | 66.4 | — | — | — | — | — | |
| VG-LAWVisual Backbone=ViT-B, Textual Backbone=BERT2024.04 | 65.36 | — | — | — | — | — | |
| SeqTRExternal Datasets=Visual Genome (5.4M) & Flickr30k-entities (158K)2024.04 | 64.69 | — | — | — | — | — | |
| VLTVisual Backbone=Swin-B, Textual Backbone=BERT2024.04 | 63.49 | — | — | — | — | — | |
| LAVTVisual Backbone=Swin-B, Textual Backbone=BERT2024.04 | 63.34 | — | — | — | — | — | |
| CM-MaskSDVisual Backbone=CLIP-ViT-B, Textual Backbone=CLIP2024.04 | 62.67 | — | — | — | — | — | |
| CRISBackbone=ResNet-1012021.11 | 59.87 | — | — | — | — | — | |
| CRISVisual Backbone=ResNet-101, Textual Backbone=CLIP2024.04 | 59.87 | — | — | — | — | — | |
| CRISBackbone=ResNet-502021.11 | 59.35 | — | — | — | — | — | |
| Grounded-SAM2024.04 | 59.14 | — | — | — | — | — | |
| VCM (trained on coco)Architecture=VCM, Training Data=COCO combined2021.05 | 57.51 | — | — | — | — | — | |
| ACM (trained on coco)Architecture=ACM, Training Data=COCO combined2021.05 | 57.49 | — | — | — | — | — | |
| LTSBackbone=DarkNet-532021.11 | 54.4 | — | — | — | — | — | |
| VLTBackbone=DarkNet-532021.11 | 52.99 | — | — | — | — | — | |
| ACMArchitecture=ACM2021.05 | 51.93 | — | — | — | — | — | |
| EFNDenseCRF=false2021.04 | 51.93 | — | — | — | — | — | |
| VCMArchitecture=VCM2021.05 | 51.22 | — | — | — | — | — | |
| CGANBackbone=DarkNet-532021.11 | 51.01 | — | — | — | — | — | |
| BUSNetDenseCRF=false2021.04 | 49.98 | — | — | — | — | — | |
| SHNetResolution=448x448, DenseCRF=false, Backbone=DeepLabv3+2021.04 | 49.9 | — | — | — | — | — | |
| TGFEfeature exchange rounds=32021.05 | 49.89 | — | — | — | — | — | |
| MCNBackbone=DarkNet-532021.11 | 49.22 | — | — | — | — | — | |
| CMPCPost-processing=DenseCRF2021.05 | 49.05 | — | — | — | — | — | |
| CMPCSource=CVPR 20202021.05 | 49.05 | — | — | — | — | — | |
| CMPCDenseCRF=true2021.04 | 49.05 | — | — | — | — | — | |
| LSCMPost-processing=DenseCRF2021.05 | 48.05 | — | — | — | — | — | |
| LSCMDenseCRF=true2021.04 | 48.05 | — | — | — | — | — | |
| BRINetPost-processing=DenseCRF2021.05 | 48.04 | — | — | — | — | — | |
| BRINDenseCRF=true2021.04 | 48.04 | — | — | — | — | — | |
| MAttNetBackbone=ResNet-1012021.11 | 47.64 | — | — | — | — | — | |
| Hu et al.Supervision=Full2024.04 | 47.57 | — | — | — | — | — | |
| NMTreeBackbone=ResNet-1012021.11 | 46.59 | — | — | — | — | — | |
| CGAN2021.05 | 46.54 | — | — | — | — | — | |
| STEP2021.05 | 46.4 | — | — | — | — | — | |
| STEP2021.05 | 46.4 | — | — | — | — | — | |
| STEPDenseCRF=true2021.04 | 46.4 | — | — | — | — | — | |
| Lang2SegBackbone=ResNet-1012021.11 | 46.37 | — | — | — | — | — | |
| CAC2021.05 | 44.32 | — | — | — | — | — | |
| SHNetResolution=320x320, DenseCRF=false, Backbone=DeepLabv3+2021.04 | 43.83 | — | — | — | — | — | |
| PPTSupervision=Weakly, Extra Pre-training Data=WebImage Text, ImageNet-1K [47]2024.04 | 42.97 | — | — | — | — | — | |
| CMSASet=val2021.02 | 42.84 | 44.71 | 37.26 | 29.1 | 19.02 | 5.59 | |
| ASGN2021.05 | 41.36 | — | — | — | — | — | |
| CMSA2019.04 | 39.98 | — | — | — | — | — | |
| CMSAPost-processing=DenseCRF2021.05 | 39.98 | — | — | — | — | — | |
| CMSA2021.05 | 39.98 | — | — | — | — | — | |
| CMSADenseCRF=true2021.04 | 39.98 | — | — | — | — | — | |
| CMSASet=val2021.02 | 39.98 | 40.38 | 32.56 | 24.77 | 16.1 | 4.59 | |
| KWA2019.04 | 36.92 | — | — | — | — | — | |
| KWA2021.05 | 36.92 | — | — | — | — | — | |
| KWA2021.05 | 36.92 | — | — | — | — | — | |
| KWASet=val2021.02 | 36.92 | 27.85 | 21.01 | 13.42 | 6.6 | 1.97 | |
| DMN2019.04 | 36.76 | — | — | — | — | — | |
| DMN2021.05 | 36.76 | — | — | — | — | — | |
| DMN2021.05 | 36.76 | — | — | — | — | — | |
| DMNSet=val2021.02 | 36.76 | — | — | — | — | — | |
| RRN2019.04 | 36.45 | — | — | — | — | — | |
| RRNPost-processing=DenseCRF2021.05 | 36.45 | — | — | — | — | — | |
| RRN2021.05 | 36.45 | — | — | — | — | — | |
| RRNDenseCRF=true2021.04 | 36.45 | — | — | — | — | — | |
| RRNSet=val2021.02 | 36.45 | 36 | 29.77 | 22.78 | 14.06 | 3.74 | |
| TRISSupervision=Weakly, Extra Pre-training Data=WebImage Text [45]2024.04 | 36 | — | — | — | — | — | |
| RMI2019.04 | 34.52 | — | — | — | — | — | |
| RMIPost-processing=DenseCRF2021.05 | 34.52 | — | — | — | — | — | |
| RMI2021.05 | 34.52 | — | — | — | — | — | |
| DMNSupervision=Full2024.04 | 34.52 | — | — | — | — | — | |
| RMISet=val2021.02 | 34.52 | 27.77 | 21.06 | 13.92 | 6.83 | 1.43 | |
| RMISupervision=Full2024.04 | 33.11 | — | — | — | — | — | |
| ChunkSupervision=Weakly, Extra Pre-training Data=CC, SBU, COCO, VG2024.04 | 32.88 | — | — | — | — | — | |
| CAFTZero-shot=true2026.02 | 31.39 | — | — | — | — | — | |
| TRIS + SAMSupervision=Weakly, Extra Pre-training Data=WebImage Text2024.04 | 29.84 | — | — | — | — | — | |
| ShatterSupervision=Weakly, Extra Pre-training Data=ImageNet-21K2024.04 | 28.87 | — | — | — | — | — | |
| LSTM-CNN2019.04 | 28.14 | — | — | — | — | — | |
| LSTM-CNN2021.05 | 28.14 | — | — | — | — | — | |
| LSTM-CNN2021.05 | 28.14 | — | — | — | — | — | |
| SaGZero-shot=true2026.02 | 25.89 | — | — | — | — | — | |
| ALBEFSupervision=Weakly, Extra Pre-training Data=CC [48], SBU [44], COCO [34], VG [23]2024.04 | 24.18 | — | — | — | — | — | |
| TSEGSupervision=Weakly, Extra Pre-training Data=ImageNet-21K [11]2024.04 | 22.05 | — | — | — | — | — | |
| FLAIRZero-shot=true2026.02 | 20.21 | — | — | — | — | — | |
| GroupViTSupervision=Weakly, Extra Pre-training Data=CC12M [7], YFCC [56]2024.04 | 16.84 | — | — | — | — | — | |
| MCLIPZero-shot=true2026.02 | 12.74 | — | — | — | — | — | |
| GViTZero-shot=true2026.02 | 10.68 | — | — | — | — | — |