Text-based Person Search on CUHK-PEDES (test)
80.05Rank-1DAPL
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| DAPL2024.05 | 80.05 | 92.65 | 97.48 | 70.24 | |
| BAMG*Additional information=human parsing2026.04 | 79.98 | 92.31 | 94.03 | 68.55 | |
| APTM2024.05 | 79.53 | 92.19 | 96.91 | 70.02 | |
| RaSa2024.05 | 78.64 | 92.33 | 96.67 | 69.28 | |
| Scale-TBPSVenue=Ours, Method Type=Non-CLIP based, NNN=true2026.01 | 77.91 | 91.07 | 94.49 | 70.88 | |
| GITM (Ours)Backbone=Swinv2-B, approach=uncertainty-based, module=group-wise ITM2026.04 | 77.88 | 91.05 | 94.57 | 72.44 | |
| MARS2024.07 | 77.62 | 90.63 | 94.27 | 71.41 | |
| MARSVenue=TOMM’25, Image enc.=Swin-B, Text enc.=BERT, Setup=with ALBEF backbone2026.01 | 77.62 | 90.63 | 94.27 | 71.71 | |
| CAMeLVenue=TIFS’25, Image enc.=SG-Former, Text enc.=BERT, Setup=with using external tools or ReID-Domain Pre-training2026.01 | 77.24 | 91.8 | 95.16 | 68.32 | |
| AULVenue=AAAI’24, Image enc.=Swin-B, Text enc.=BERT, Setup=with using external tools or ReID-Domain Pre-training2026.01 | 77.23 | 90.43 | 94.41 | — | |
| AUL2026.04 | 77.23 | 90.43 | 94.41 | — | |
| DiCo2026.04 | 77.21 | 91.85 | 95.63 | — | |
| CADAretrained=true2024.07 | 77.2 | 90.68 | 93.92 | 68.45 | |
| DECL2024.05 | 77.06 | 92.37 | 96.16 | 68.43 | |
| OursVenue=WACV’26, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with CLIP backbone2026.01 | 76.95 | 90.64 | 94.36 | 69.38 | |
| ITSELF2026.04 | 76.95 | 90.64 | 94.36 | 69.38 | |
| Baseline (APTM)Backbone=Swinv2-B2026.04 | 76.9 | 90.75 | 94.33 | 68.85 | |
| Scale-TBPSVenue=Ours, Method Type=Non-CLIP based, NNN=false2026.01 | 76.85 | 90.22 | 94.23 | 70.05 | |
| IRRA2024.05 | 76.82 | 91.23 | 95.58 | 65.89 | |
| APTMnumber of parameters=214M2024.04 | 76.53 | 90.04 | 94.15 | 66.91 | |
| APTMVenue=MM, Method Type=Non-CLIP based, NNN=false2026.01 | 76.53 | 90.04 | 94.15 | 66.91 | |
| APTMVenue=MM’23, Image enc.=Swin-B, Text enc.=BERT, Setup=with using external tools or ReID-Domain Pre-training2026.01 | 76.53 | 90.04 | 94.15 | 66.91 | |
| APTM2026.04 | 76.53 | 90.04 | 94.15 | 66.91 | |
| RaSaVision-language pretraining (VLP)=true, Re-ranking=false, Backbone=ALBEF2023.05 | 76.51 | 90.29 | 94.25 | 69.38 | |
| RaSaCLIP-based=false2023.08 | 76.51 | 90.29 | 94.25 | 69.38 | |
| RaSanumber of parameters=210M2024.04 | 76.51 | 90.29 | 94.25 | 69.38 | |
| RaSa2024.07 | 76.51 | 90.29 | 94.25 | 69.38 | |
| RaSaVenue=IJCAI, Method Type=Non-CLIP based, NNN=false2026.01 | 76.51 | 90.29 | 94.25 | 69.38 | |
| RaSaVenue=IJCAI’23, Image enc.=Swin-B, Text enc.=BERT, Setup=with ALBEF backbone2026.01 | 76.51 | 90.29 | 94.25 | 69.38 | |
| RaSa2026.04 | 76.51 | 90.29 | 94.25 | 69.38 | |
| ICLVenue=CVPR, Method Type=CLIP based, NNN=false2026.01 | 76.41 | 90.48 | 94.33 | 68.04 | |
| WoRAnumber of parameters=127M2024.04 | 76.38 | 89.72 | 93.49 | 67.22 | |
| WoRA-TBPSVenue=MM, Method Type=Non-CLIP based, NNN=false2026.01 | 76.38 | 89.72 | 93.49 | 67.22 | |
| RDEVenue=-, Method Type=Non-CLIP based, NNN=true2026.01 | 76.32 | 90.95 | 94.43 | 68.13 | |
| RDEVenue=CVPR, Method Type=Non-CLIP based, NNN=false2026.01 | 75.94 | 90.14 | 94.12 | 67.56 | |
| RDEVenue=CVPR’24, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with CLIP backbone2026.01 | 75.94 | 90.14 | 94.12 | 67.56 | |
| RDE2026.04 | 75.94 | 90.14 | 94.12 | 67.56 | |
| DPVenue=AAAI’24, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with using external tools or ReID-Domain Pre-training2026.01 | 75.66 | 90.59 | 94.07 | 66.58 | |
| CFAMVenue=CVPR, Method Type=CLIP based, NNN=false2026.01 | 75.6 | 90.53 | 94.36 | 67.27 | |
| CFAM(L/14)Venue=CVPR’24, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with using external tools or ReID-Domain Pre-training2026.01 | 75.6 | 90.53 | 94.36 | 67.27 | |
| APTM (Re-implementation)number of parameters=214M, note=Baseline*2024.04 | 75.42 | 88.86 | 92.77 | 66.61 | |
| PLOTVenue=ECCV’24, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with CLIP backbone2026.01 | 75.28 | 90.42 | 94.12 | — | |
| SAP-SAMVenue=MM’24, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with using external tools or ReID-Domain Pre-training2026.01 | 75.05 | 89.93 | 93.73 | — | |
| CRUEVenue=TOMM’25, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with CLIP backbone2026.01 | 74.91 | 90.11 | 93.9 | 68.93 | |
| MUMVenue=AAAI’24, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with CLIP backbone2026.01 | 74.25 | 89.83 | 93.58 | 66.15 | |
| BiLMaVenue=ICCV’23, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with CLIP backbone2026.01 | 74.03 | 89.59 | 93.62 | 66.57 | |
| SAMC2026.04 | 74.03 | 89.18 | 93.31 | 68.42 | |
| TBPS-CLIPCLIP-based=true, Backbone=ViT-B/162023.08 | 73.54 | 88.19 | 92.35 | 65.38 | |
| TBPS-CLIPnumber of parameters=149M2024.04 | 73.54 | 88.19 | 92.35 | 65.38 | |
| TBPS-CLIP2024.07 | 73.54 | 88.19 | 92.35 | 65.38 | |
| TBPS-CLIPVenue=AAAI, Method Type=CLIP based, NNN=false2026.01 | 73.54 | 88.19 | 92.35 | 65.38 | |
| TBPSCLIPVenue=AAAI’24, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with CLIP backbone2026.01 | 73.54 | 88.19 | 92.35 | 65.38 | |
| TBPS-CLIP2026.04 | 73.54 | 88.19 | 92.35 | 65.38 | |
| IRRACLIP-based=true, Backbone=ViT-B/162023.08 | 73.38 | 89.93 | 93.71 | 66.13 | |
| IRRAnumber of parameters=194M2024.04 | 73.38 | 89.93 | 93.71 | 66.13 | |
| IRRA2024.07 | 73.38 | 89.93 | 93.71 | 66.13 | |
| IRRAVenue=CVPR, Method Type=CLIP based, NNN=false2026.01 | 73.38 | 89.93 | 93.71 | 66.13 | |
| IRRAVenue=CVPR’23, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with CLIP backbone2026.01 | 73.38 | 89.93 | 93.71 | 66.13 | |
| IRRA2026.04 | 73.38 | 89.93 | 93.71 | 66.13 | |
| MACF2026.04 | 73.33 | 88.57 | 93.02 | — | |
| CFAM (B/16)Venue=CVPR’24, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with using external tools or ReID-Domain Pre-training2026.01 | 72.87 | 88.61 | 92.87 | 64.92 | |
| Simplified TBPS-CLIPCLIP-based=true, Backbone=ViT-B/162023.08 | 72.66 | 88.14 | 92.72 | 64.97 | |
| DM-AdapterVenue=AAAI’25, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with CLIP backbone2026.01 | 72.17 | 88.74 | 92.85 | 64.33 | |
| VGSG2024.04 | 71.38 | 86.75 | 91.86 | 67.91 | |
| VGSG2026.04 | 71.38 | 86.75 | 91.86 | 67.91 | |
| TP-TPSCLIP-based=true2023.08 | 70.16 | 86.1 | 90.98 | 66.32 | |
| CSKTVenue=ICASSP’24, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with CLIP backbone2026.01 | 69.7 | 86.92 | 91.8 | 62.74 | |
| CFineVision-language pretraining (VLP)=true, Re-ranking=false2023.05 | 69.57 | 85.93 | 91.15 | — | |
| CFineCLIP-based=true2023.08 | 69.57 | 85.93 | 91.15 | — | |
| CFine2024.04 | 69.57 | 85.93 | 91.15 | — | |
| CFine2024.07 | 69.57 | 85.93 | 91.15 | — | |
| CFineVenue=TIP, Method Type=CLIP based, NNN=false2026.01 | 69.57 | 85.93 | 91.15 | — | |
| CFineVenue=TIP’23, Image enc.=CLIP-ViT, Text enc.=BERT, Setup=with CLIP backbone2026.01 | 69.57 | 85.93 | 91.15 | — | |
| CFine2026.04 | 69.57 | 85.93 | 91.15 | — | |
| TBPS-CLIPCLIP-based=true, Backbone=ViT-B/322023.08 | 69.54 | 86.99 | 91.24 | 61.57 | |
| CLIPVenue=PMLR, Method Type=CLIP based, NNN=false2026.01 | 68.55 | 86.32 | 91.85 | 61.25 | |
| UniPTVenue=ICCV’23, Image enc.=ViT-B/16, Text enc.=BERT, Setup=with using external tools or ReID-Domain Pre-training2026.01 | 68.5 | 84.67 | 90.38 | — | |
| TransTPS2026.04 | 68.23 | 86.37 | 91.65 | — | |
| CLIPbackbone=ViT-B/16, mode=fine-tuned2024.07 | 68.19 | 86.47 | 91.47 | 61.12 | |
| VGSGCAM=false, FLOPs=6.79G, Inference=5.3ms2023.11 | 67.1 | — | — | — | |
| CLIP (ViT-B/16)Venue=ICML’21, Image enc.=CLIP-ViT, Text enc.=CLIP-Xformer, Setup=with CLIP backbone2026.01 | 65.73 | 86.39 | 92.01 | 61.97 | |
| IVTVision-language pretraining (VLP)=true, Re-ranking=false2023.05 | 65.59 | 83.11 | 89.21 | — | |
| IVTCLIP-based=false2023.08 | 65.59 | 83.11 | 89.21 | 60.66 | |
| CLIPCLIP-based=true, Backbone=ViT-B/16, Protocol=Baseline2023.08 | 65.37 | 85.83 | 91.59 | 59.42 | |
| LGURVision-language pretraining (VLP)=false, Re-ranking=false2023.05 | 65.25 | 83.12 | 89 | — | |
| LGURCLIP-based=false2023.08 | 65.25 | 83.12 | 89 | — | |
| LGUR2024.04 | 65.25 | 83.12 | 89 | — | |
| LGUR2026.04 | 65.25 | 83.12 | 89 | — | |
| C2A2Vision-language pretraining (VLP)=false, Re-ranking=false2023.05 | 64.82 | 83.54 | 89.77 | — | |
| AXM-NetCLIP-based=false2023.08 | 64.44 | 80.52 | 86.77 | 58.73 | |
| AXM-Net2024.04 | 64.44 | 80.52 | 86.77 | 58.73 | |
| CAIBCVision-language pretraining (VLP)=false, Re-ranking=false2023.05 | 64.43 | 82.87 | 88.37 | — | |
| CAIBCCLIP-based=false2023.08 | 64.43 | 82.87 | 88.37 | — | |
| CAIBC2024.04 | 64.43 | 82.87 | 88.37 | — | |
| CAIBC2026.04 | 64.43 | 82.87 | 88.37 | — | |
| BLIPmode=fine-tuned2024.07 | 64.36 | 83.36 | 88.78 | 58.18 | |
| TIPCBVision-language pretraining (VLP)=false, Re-ranking=false2023.05 | 64.26 | 83.19 | 89.1 | — | |
| SAFVision-language pretraining (VLP)=false, Re-ranking=false2023.05 | 64.13 | 82.62 | 88.4 | 58.61 | |
| SAFCLIP-based=false2023.08 | 64.13 | 82.62 | 88.4 | 58.61 | |
| PSLDVision-language pretraining (VLP)=true, Re-ranking=false2023.05 | 64.08 | 81.73 | 88.19 | 60.08 |