Crowd Counting on ShanghaiTech Part A (test)
46MAEFSCA-Net+CSRNET(M)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| FSCA-Net+CSRNET(M)Backbone=CSRNet, Training Protocol=Joint Training2026.02 | 46 | 87.6 | — | |
| FSCA-Net+BL(M)Backbone=BL, Training Protocol=Joint Training2026.02 | 46.1 | 78.1 | — | |
| PET2026.02 | 47.4 | 75 | — | |
| CrowDiff2026.02 | 47.4 | 75 | — | |
| APGCC2026.02 | 48.8 | 76.7 | — | |
| APGCCTraining=Single-scene2026.03 | 48.8 | 76.7 | — | |
| FSCA-Net+MAN(M)Backbone=MAN, Training Protocol=Joint Training2026.02 | 49.3 | 78.2 | — | |
| PETTraining=Single-scene2026.03 | 49.3 | 78.7 | — | |
| FSCA-Net+DM(M)Backbone=DM, Training Protocol=Joint Training2026.02 | 49.9 | 85.3 | — | |
| DiffusionLocTraining=Single-scene2026.03 | 50.3 | 79.1 | — | |
| P2R2026.02 | 51 | 79.7 | — | |
| PSDNN + Crowd HatType=Detection-Based2023.08 | 51.2 | — | 81.9 | |
| CrowdHat2026.02 | 51.2 | 81.9 | — | |
| FGENetVenue=This Paper2024.01 | 51.66 | 85 | — | |
| FGENetLocalization=true, Architecture=CNN2024.05 | 51.66 | 85 | — | |
| VMambaCCLocalization=true, Architecture=Mamba2024.05 | 51.87 | 81.3 | — | |
| CLIP-EBCType=Classification-based (C), Backbone=ViT/B-162024.03 | 52.5 | — | 85.9 | |
| P2P-NetType=Localization-Based2023.08 | 52.7 | — | 85.1 | |
| P2PNetLocalization=true, Architecture=CNN2024.05 | 52.7 | 85.1 | — | |
| P2PNetType=Regression-based (R)2024.03 | 52.7 | — | 85 | |
| P2PNetYear=2024, Backbone=VGG162026.01 | 52.74 | 85.06 | — | |
| SDNet + Crowd HatType=Detection-Based2023.08 | 53.4 | — | 87.2 | |
| CLIP-EBCType=Classification-based (C), Backbone=ResNet502024.03 | 54 | — | 83.2 | |
| STEERERType=Regression-based (R)2024.03 | 54.5 | — | 86.9 | |
| STEERER2026.02 | 54.5 | 86.9 | — | |
| STEERERYear=2024, Backbone=HRNet2026.01 | 54.5 | 86.9 | — | |
| UEPNetType=Classification-based (C)2024.03 | 54.6 | — | 91.2 | |
| GramformerYear=2024, Backbone=VGG192026.01 | 54.7 | 87.1 | — | |
| GauNet (ResNet-50)Backbone=ResNet-502022.06 | 54.8 | 89.1 | — | |
| GauNetVenue=CVPR'222024.01 | 54.8 | 89.1 | — | |
| GauNetLocalization=false, Architecture=CNN2024.05 | 54.8 | 89.1 | — | |
| LoViTCrowdLocalization=false, Architecture=Transformer2024.05 | 54.8 | 80.9 | — | |
| GauNet2026.02 | 54.8 | 89.1 | — | |
| GauNetTraining=Single-scene2026.03 | 54.8 | 89.1 | — | |
| RepSFNetYear=2025, Backbone=RepLKViT2026.01 | 54.9 | 87.6 | — | |
| SDNetVenue=ICCV’212022.06 | 55 | 92.7 | — | |
| ADSCNetVenue=CVPR’202022.06 | 55.4 | 97.7 | — | |
| ADSCNetType=Density-Based2023.08 | 55.4 | — | 97.7 | |
| ADSCNET2026.02 | 55.4 | 97.7 | — | |
| P2Rlabeled Pct.=40%2026.03 | 55.6 | 95 | — | |
| P2PNetVenue=ICCV'212024.01 | 55.73 | 89.2 | — | |
| XMasklabeled Pct.=40%2026.03 | 56.2 | 93.5 | — | |
| AMSNetVenue=ECCV’202022.06 | 56.7 | 93.4 | — | |
| AMSNetYear=2020, Backbone=VGG192026.01 | 56.7 | 93.4 | — | |
| MANType=Density-Based2023.08 | 56.8 | — | 90.3 | |
| MANVenue=CVPR'222024.01 | 56.8 | 90.3 | — | |
| CLTRType=Localization-Based2023.08 | 56.9 | — | 95.2 | |
| CrowdFormerType=Density-Based2023.08 | 56.9 | — | 97.4 | |
| CLTRVenue=ECCV'232024.01 | 56.9 | 95.2 | — | |
| CLTRLocalization=true, Architecture=Transformer2024.05 | 56.9 | 95.2 | — | |
| CLTRType=Regression-based (R)2024.03 | 56.9 | — | 95.2 | |
| S3Venue=IJCAI'212024.01 | 57 | 96 | — | |
| GLossVenue=IJCAI'212024.01 | 57.3 | 90.7 | — | |
| SCPNetYear=2023, Backbone=HRNet2026.01 | 57.3 | 102.1 | — | |
| GGANetYear=2023, Backbone=GGANet2026.01 | 57.4 | 110.7 | — | |
| ChfLVenue=CVPR'222024.01 | 57.5 | 94.3 | — | |
| ChfL2026.02 | 57.5 | 94.3 | — | |
| ChflYear=2022, Backbone=VGG192026.01 | 57.5 | 94.3 | — | |
| M-SFANet+M-SegNetinference_strategy=average prediction2020.03 | 57.55 | — | 94.48 | |
| M-SFANet+M-SegNetYear=2021, Backbone=VGG16-BN2026.01 | 57.55 | 94.48 | — | |
| ASNetVenue=CVPR'202024.01 | 57.78 | 90.13 | — | |
| ASNetLocalization=false, Architecture=CNN2024.05 | 57.78 | 90.13 | — | |
| ASNPE2026.02 | 57.8 | 97.1 | — | |
| S-DCNet2019.08 | 58.3 | 95 | — | |
| S-DCNet2020.03 | 58.3 | — | 95 | |
| S-DCNetType=Classification-based (C)2024.03 | 58.3 | — | 95 | |
| GauNet (SANet)Backbone=SANet2022.06 | 59.2 | 95.4 | — | |
| CHS-NetVenue=ICASSP'232024.01 | 59.2 | 97.8 | — | |
| SANet+SPANet2019.09 | 59.4 | 92.5 | — | |
| SANet+SPANet2020.03 | 59.4 | — | 92.5 | |
| M-SFANet2020.03 | 59.69 | — | 95.66 | |
| M-SFANetYear=2021, Backbone=VGG16-BN2026.01 | 59.69 | 95.66 | — | |
| DM-Count2020.09 | 59.7 | — | 95.7 | |
| DM-CountVenue=NeurIPS’202022.06 | 59.7 | 95.7 | — | |
| DM-CountMult-Adds=853.70 G2022.02 | 59.7 | — | 95.7 | |
| DMCountType=Regression-based (R)2024.03 | 59.7 | — | 95.7 | |
| M-SFANET2026.02 | 59.7 | 95.7 | — | |
| DMTraining Protocol=Separate Training2026.02 | 59.7 | 95.7 | — | |
| DM-CountYear=2020, Backbone=VGG192026.01 | 59.7 | 95.7 | — | |
| DM-CountTraining=Single-scene2026.03 | 59.7 | 95.7 | — | |
| SFANet2020.03 | 59.8 | — | 99.3 | |
| S-DCNetYear=2022, Backbone=VGG162026.01 | 59.8 | 100 | — | |
| M-SFANetAblation=w/o skip connections2020.03 | 60.07 | — | 99.47 | |
| MBTTBF-SCFB2020.03 | 60.2 | — | 94.1 | |
| LSC-CNN + Crowd HatType=Detection-Based2023.08 | 60.2 | — | 95.7 | |
| M-SegNet2020.03 | 60.55 | — | 100.8 | |
| M-SegNetYear=2021, Backbone=VGG16-BN2026.01 | 60.55 | 100.8 | — | |
| SRRNetYear=2023, Backbone=HRNet2026.01 | 60.8 | 108.3 | — | |
| GauNet (CSRNet)Backbone=CSRNet2022.06 | 61.2 | 97.8 | — | |
| Topo-CountType=Localization-Based2023.08 | 61.2 | — | 104.6 | |
| RPNNet2026.02 | 61.2 | 96.9 | — | |
| M-SFANetAblation=w/o ASPP module2020.03 | 61.25 | — | 102.37 | |
| GLossVenue=CVPR’212022.06 | 61.3 | 95.4 | — | |
| GLType=Localization-Based2023.08 | 61.3 | — | 95.4 | |
| GL2026.02 | 61.3 | 95.4 | — | |
| GLossTraining=Single-scene2026.03 | 61.3 | 95.4 | — | |
| SPN2019.08 | 61.7 | 99.5 | — | |
| MNAVenue=NeurIPS’202022.06 | 61.9 | 99.6 | — | |
| MNA2026.02 | 61.9 | 99.6 | — | |
| FusionCountMult-Adds=815.00 G2022.02 | 62.2 | — | 101.2 |