Crowd Counting on ShanghaiTech Part A
52.74MAEP2PNet
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| P2PNet2021.07 | 52.74 | 85.06 | — | |
| P2PNetYear=2024, Backbone=VGG162026.01 | 52.74 | 85.06 | — | |
| HCA-dApproach=Classification approach2023.10 | 53.7 | — | 87.8 | |
| STEERERYear=2024, Backbone=HRNet2026.01 | 54.5 | 86.9 | — | |
| HCA-mulApproach=Classification approach2023.10 | 54.7 | — | 91.6 | |
| GramformerYear=2024, Backbone=VGG192026.01 | 54.7 | 87.1 | — | |
| RepSFNetYear=2025, Backbone=RepLKViT2026.01 | 54.9 | 87.6 | — | |
| SDANetYear=2024, Supervision Type=FS, Parameters (M)=56.52025.12 | 54.9 | — | 90.4 | |
| ADSCNetVenue=CVPR'202021.07 | 55.4 | 97.7 | — | |
| HCA-addApproach=Classification approach2023.10 | 55.9 | — | 92.8 | |
| AMSNetVenue=ECCV'202021.07 | 56.7 | 93.4 | — | |
| SGANetCurriculum Loss=true2019.11 | 57.6 | — | 101.1 | |
| ASNetVenue=CVPR'202021.07 | 57.78 | 90.13 | — | |
| SGANet2019.11 | 58 | — | 100.4 | |
| CLSApproach=Classification approach2023.10 | 58.2 | — | 96.7 | |
| S-DCNetVenue=ICCV'192021.07 | 58.3 | 95 | — | |
| DLPTNetYear=2024, Supervision Type=FS, Parameters (M)=110.92025.12 | 58.4 | — | 95 | |
| RAQNetYear=2024, Supervision Type=FS, Parameters (M)=42.772025.12 | 59 | — | 101.2 | |
| PaDNetApproach=Density map regression2023.10 | 59.2 | — | 98.1 | |
| DHMoEYear=2025, Supervision Type=FS, Parameters (M)=5.862025.12 | 59.2 | — | 96.1 | |
| RANet2019.11 | 59.4 | — | 102 | |
| SPANet2019.11 | 59.4 | — | 92.5 | |
| SANet + SPANetVenue=ICCV'192021.07 | 59.4 | 92.5 | — | |
| DM-CountVenue=NeurIPS'202021.07 | 59.7 | 95.7 | — | |
| OTApproach=Density map regression2023.10 | 59.7 | — | 95.7 | |
| Inception-v3status=modified2019.11 | 60.1 | — | 105 | |
| DC-regressionApproach=Local count regression2023.10 | 60.7 | — | 101 | |
| SALCrowdVenue=ACM MM'24, Labeled Percentage=40%2025.03 | 60.7 | — | 97.2 | |
| TMTBVenue=Ours, Labeled Percentage=40%2025.03 | 60.8 | — | 99 | |
| SRRNetYear=2023, Supervision Type=FS, Parameters (M)=66.142025.12 | 60.8 | — | 103 | |
| RPNetmethod_type=single-class crowd counting2022.01 | 61.2 | — | — | |
| GLFmethod_type=single-class crowd counting2022.01 | 61.3 | — | — | |
| GLApproach=Density map regression2023.10 | 61.3 | — | 95.4 | |
| AMRNetVenue=ECCV'202021.07 | 61.59 | 98.36 | — | |
| MNAApproach=Density map regression2023.10 | 61.9 | — | 99.6 | |
| MRC-CrowdVenue=T-CSVT'24, Labeled Percentage=40%2025.03 | 62.1 | — | 95.5 | |
| CANNet2019.11 | 62.3 | — | 100 | |
| CANVenue=CVPR'192021.07 | 62.3 | 100 | — | |
| Bayesian+Venue=ICCV'192021.07 | 62.8 | 101.8 | — | |
| BLApproach=Density map regression2023.10 | 62.8 | — | 101.8 | |
| LMSNetYear=2024, Supervision Type=FS, Parameters (M)=0.732025.12 | 62.9 | — | 108.4 | |
| ADCrowdNet(AMG-bAttn-DME)Architecture=AMG-bAttn-DME2018.11 | 63.2 | 98.9 | — | |
| SDANetVenue=AAAI'202021.07 | 63.6 | 101.8 | — | |
| ANF2019.11 | 63.9 | — | 99.4 | |
| DRCNApproach=Density map regression2023.10 | 64 | — | 98.4 | |
| DADNet2019.11 | 64.2 | — | 99.9 | |
| TEDNet2019.11 | 64.2 | — | 109.1 | |
| DKDYear=2023, Backbone=VGG192026.01 | 64.4 | 103 | — | |
| TCFormerYear=2025, Supervision Type=NF, Parameters (M)=5.522025.12 | 64.541 | — | 110.052 | |
| Wan et al.2019.11 | 64.7 | — | 97.1 | |
| SFCNYear=2021, Supervision Type=FS, Parameters (M)=38.62025.12 | 64.8 | — | 107.5 | |
| DeepCount2019.08 | 65.2 | — | 112.5 | |
| RegressionApproach=Local count regression2023.10 | 65.4 | — | 103.3 | |
| TMTBVenue=Ours, Labeled Percentage=10%2025.03 | 65.7 | — | 110.4 | |
| PSDDN+2019.08 | 65.9 | — | 112.3 | |
| ADCrowdNet(AMG-DME)Architecture=AMG-DME2018.11 | 66.1 | 102.1 | — | |
| CSFNetYear=2025, Backbone=VGG192026.01 | 66.1 | 103.2 | — | |
| SANet2019.11 | 67 | — | 104.5 | |
| SANet2019.08 | 67 | — | 104.5 | |
| MRC-CrowdVenue=T-CSVT'24, Labeled Percentage=10%2025.03 | 67.3 | — | 106.8 | |
| DACountVenue=ACM MM'22, Labeled Percentage=40%2025.03 | 67.5 | — | 110.7 | |
| CSRNet2018.11 | 68.2 | 115 | — | |
| CSRNet2019.11 | 68.2 | — | 115 | |
| CSRNet2019.08 | 68.2 | — | 115 | |
| CSRNetmethod_type=single-class crowd counting2022.01 | 68.2 | — | — | |
| CSRNetApproach=Density map regression2023.10 | 68.2 | — | 115 | |
| ADCrowdNet(DME)Architecture=DME2018.11 | 68.5 | 107.5 | — | |
| ic-CNN2019.08 | 68.5 | — | 116.2 | |
| SUAVenue=ICCV'21, Labeled Percentage=40%2025.03 | 68.5 | — | 121.9 | |
| SALCrowdVenue=ACM MM'24, Labeled Percentage=10%2025.03 | 69.7 | — | 114.5 | |
| Zhang et al.Year=2024, Supervision Type=NF2025.12 | 69.7 | — | 114.5 | |
| ic-CNNTraining Time=10 hrs, Number of Parameters=7.9 x 10^6, Hardware=Nvidia GTX 1080 TI2018.07 | 69.8 | — | — | |
| ImprovedCSRNetYear=2025, Backbone=VGG192026.01 | 70.29 | 116.6 | — | |
| OT-MVenue=CVPR'23, Labeled Percentage=40%2025.03 | 70.7 | — | 114.5 | |
| CUVenue=ICCV'23, Labeled Percentage=10%2025.03 | 70.8 | — | 116.6 | |
| ADCrowdNet(AMG-attn-DME)Architecture=AMG-attn-DME2018.11 | 70.9 | 115.2 | — | |
| Multi-task trainingranking_data_source=Query-by-example2018.03 | 72 | 106.6 | — | |
| TMTBVenue=Ours, Labeled Percentage=5%2025.03 | 72.4 | — | 126.4 | |
| PDDNetYear=2024, Supervision Type=FS, Parameters (M)=1.12025.12 | 72.6 | — | 112.2 | |
| CP-CNNTraining Time=unknown, Number of Parameters=6.3 x 10^72018.07 | 73.6 | — | — | |
| CP-CNN2017.08 | 73.6 | 106.4 | — | |
| Sindagi et al.2018.03 | 73.6 | 106.4 | — | |
| Multi-task trainingranking_data_source=Keyword2018.03 | 73.6 | 112 | — | |
| CP-CNN2018.11 | 73.6 | 106.4 | — | |
| CP-CNN2019.08 | 73.6 | — | 106.4 | |
| CP-CNNmethod_type=single-class crowd counting2022.01 | 73.6 | — | — | |
| SAFECountmethod_type=few-shot counting2022.01 | 73.7 | — | — | |
| MRC-CrowdVenue=T-CSVT'24, Labeled Percentage=5%2025.03 | 74.8 | — | 117.3 | |
| DACountVenue=ACM MM'22, Labeled Percentage=10%2025.03 | 74.9 | — | 115.5 | |
| TinyCountYear=2024, Supervision Type=FS, Parameters (M)=0.062025.12 | 78.2 | — | 120.8 | |
| OT-MVenue=CVPR'23, Labeled Percentage=10%2025.03 | 80.1 | — | 118.5 | |
| OT-MYear=2023, Supervision Type=NF2025.12 | 80.1 | — | 118.5 | |
| MRLYear=2023, Supervision Type=NF2025.12 | 80.2 | — | 125.6 | |
| MTCPVenue=T-NNLS'23, Labeled Percentage=10%2025.03 | 81.3 | — | 130.5 | |
| DACountYear=2022, Supervision Type=NF2025.12 | 82.5 | — | 123.2 | |
| OT-MVenue=CVPR'23, Labeled Percentage=5%2025.03 | 83.7 | — | 133.3 | |
| RepMobileNetYear=2024, Supervision Type=FS, Parameters (M)=3.412025.12 | 84.2 | — | 127.5 | |
| DACountVenue=ACM MM'22, Labeled Percentage=5%2025.03 | 85.4 | — | 134.5 | |
| LMSFFNetYear=2024, Supervision Type=FS, Parameters (M)=2.882025.12 | 85.9 | — | 139.9 | |
| L2RVenue=CVPR'18, Labeled Percentage=40%2025.03 | 86.5 | — | 148.2 |