Car Object Counting on CARPK (test)
2.12MAEHLCNN
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| HLCNN2025.07 | 2.12 | 3.02 | — | — | — | — | — | |
| HLCNNMethod category=Single-class car counting method2022.01 | 2.12 | 3.02 | — | — | — | — | — | |
| ODCi=989, Ca=ALL2019.05 | 2.94 | 3.94 | 96.67 | — | — | — | — | |
| CLIP-EBC-CARdomain adaptation=VPT, parameters=frozen, text prompts=modified2025.07 | 4.01 | 6.02 | — | — | — | — | — | |
| ODCi=5, Ca=ALL2019.05 | 4.38 | 5.58 | 91.97 | — | — | — | — | |
| A-Simple-ButTraining Status=Training-Free, Exemplar Usage=Utilize exemplars2026.01 | 4.39 | 5.7 | — | — | — | — | — | |
| SAFECount2025.07 | 5.33 | 7.04 | — | — | — | — | — | |
| SAFECountMethod category=Few-shot counting method2022.01 | 5.33 | 7.04 | — | — | — | — | — | |
| CounTR2025.07 | 5.75 | 7.45 | — | — | — | — | — | |
| CounTRYear=2022, Method=Regression, Type=Generic2022.08 | 5.75 | 7.45 | — | — | — | — | — | |
| BMNet+2025.07 | 5.76 | 7.83 | — | — | — | — | — | |
| BMNet+fine-tuned=true2022.03 | 5.76 | — | — | 7.83 | — | — | — | |
| BMNet+Fine-Tuned=true, Reference protocol=Reference-based2022.05 | 5.76 | 7.83 | — | — | — | — | — | |
| BMNet+Year=2022, Method=Regression, Type=Generic2022.08 | 5.76 | 7.83 | — | — | — | — | — | |
| BMNet+Training=Yes, Prompt=box2023.06 | 5.76 | 7.83 | — | — | — | — | — | |
| ODCi=5, Ca=502019.05 | 6.02 | 8.23 | 87.9 | — | — | — | — | |
| VLCounter2025.07 | 6.46 | 8.68 | — | — | — | — | — | |
| VLCounterEvaluation Protocol=zero-shot, Stage=1, Class Name=Checkmark, Train Dataset=FSC1472023.12 | 6.46 | 8.68 | — | — | — | — | — | |
| Our full approachbackbone=RetinaNet, components=Soft-IoU layer, EM-Merger unit2019.04 | 6.77 | 8.52 | — | — | — | — | — | |
| PDEMMethod category=Multi-class counting method2022.01 | 6.77 | 8.52 | — | — | — | — | — | |
| PDEMYear=2019, Method=Detection, Type=Specific2022.08 | 6.77 | 8.52 | — | — | — | — | — | |
| GMNextra_training_data=ILSVRC video dataset2021.04 | 7.48 | 9.9 | — | — | — | — | — | |
| GMNMethod category=Few-shot counting method2022.01 | 7.48 | 9.9 | — | — | — | — | — | |
| GMNYear=2021, Method=Regression, Type=Generic2022.08 | 7.48 | 9.9 | — | — | — | — | — | |
| GMNTraining=Yes, Prompt=box2023.06 | 7.48 | 9.9 | — | — | — | — | — | |
| BMNetfine-tuned=true2022.03 | 8.05 | — | — | 9.7 | — | — | — | |
| BMNetFine-Tuned=true, Reference protocol=Reference-based2022.05 | 8.05 | 9.7 | — | — | — | — | — | |
| CounTXzero-shot=true2025.02 | 8.13 | 10.87 | — | — | — | — | — | |
| ABACUS-3BSupervision=Zero-shot/MLLM (Z)2026.06 | 8.41 | 10.84 | — | — | — | — | — | |
| T2ICountzero-shot=true2025.02 | 8.61 | 13.47 | — | — | — | — | — | |
| T2ICountSupervision=Point-level (P)2026.06 | 8.61 | 13.47 | — | — | — | — | — | |
| RCCFine-Tuned=true, Reference protocol=Reference-less2022.05 | 9.21 | 11.33 | — | — | — | — | — | |
| RCCTraining Status=Rely on training, Exemplar Usage=Prior-Free, Fine-tuned on CARPK=true2026.01 | 9.21 | 11.33 | — | — | — | — | — | |
| ValidCounteTraining Status=Training-Free, Exemplar Usage=Utilize exemplars2026.01 | 9.36 | 12.86 | — | — | 0.17 | 1.41 | — | |
| SAVETraining Status=Rely on training, Exemplar Usage=Prior-Free2026.01 | 9.56 | 19.04 | — | — | — | — | — | |
| SSDBackBone=ResNet-502024.05 | 9.58 | 12.15 | — | — | — | — | — | |
| TFCounterTraining Status=Training-Free, Exemplar Usage=Utilize exemplars2026.01 | 9.71 | 12.44 | — | — | — | — | — | |
| OmniCountTraining Status=Training-Free, Exemplar Usage=Utilize exemplars, Multi-Class=true2026.01 | 9.92 | 12.15 | — | — | 0.23 | 2.11 | — | |
| LOCASetting=Few-shot2022.11 | 9.97 | 12.51 | — | — | — | — | — | |
| LOCABackBone=ResNet-502024.05 | 9.97 | 12.51 | — | — | — | — | — | |
| OCCAM-STraining Status=Rely on training, Exemplar Usage=Prior-Free, Multi-Class=true2026.01 | 10.06 | 13.81 | — | — | 0.13 | 1.41 | — | |
| WS-COC-7BSupervision=Image-level (I)2026.06 | 10.39 | 15.83 | — | — | — | — | — | |
| BMNet+fine-tuned=false2022.03 | 10.44 | — | — | 13.77 | — | — | — | |
| BMNet+Fine-Tuned=false, Reference protocol=Reference-based2022.05 | 10.44 | 13.77 | — | — | — | — | — | |
| BMNet+Setting=Few-shot2022.11 | 10.44 | 13.77 | — | — | — | — | — | |
| BMNet+Evaluation Protocol=few-shot, Stage=1, Class Name=X, Train Dataset=FSC1472023.12 | 10.44 | 13.77 | — | — | — | — | — | |
| BMNetBackBone=ResNet-502024.05 | 10.44 | 13.77 | — | — | — | — | — | |
| VA-Countzero-shot=true2025.02 | 10.63 | 13.2 | — | — | — | — | — | |
| GCA-SUNetCategory=EFC, Training Dataset=FSC-1472024.09 | 10.96 | 13.95 | — | — | — | — | — | |
| GCA-SUNTraining Status=Rely on training, Exemplar Usage=Prior-Free2026.01 | 10.96 | 13.95 | — | — | — | — | — | |
| OursTraining=No, Prompt=box2023.06 | 10.97 | 14.24 | — | — | 0.48 | 3.7 | — | |
| TFOCTraining Status=Training-Free, Exemplar Usage=Utilize exemplars2026.01 | 10.97 | 14.24 | — | — | 0.48 | 3.7 | — | |
| OursTraining=No, Prompt=point2023.06 | 11.01 | 14.34 | — | — | 0.51 | 3.89 | — | |
| CounTRCategory=CAC, Training Dataset=FSC-147, Venue=BMVC'222024.09 | 11.52 | 14.56 | — | — | — | — | — | |
| PFODCi=5, Ca=502019.05 | 11.54 | 16.76 | 83.73 | — | — | — | — | |
| CLIP-Countzero-shot=true2025.02 | 11.96 | 16.61 | — | — | — | — | — | |
| RCCFine-Tuned=false, Reference protocol=Reference-less2022.05 | 12.31 | 15.4 | — | — | — | — | — | |
| BMNetfine-tuned=false2022.03 | 14.61 | — | — | 24.6 | — | — | — | |
| BMNetFine-Tuned=false, Reference protocol=Reference-based2022.05 | 14.61 | 24.6 | — | — | — | — | — | |
| BMNetEvaluation Protocol=few-shot, Stage=1, Class Name=X, Train Dataset=FSC1472023.12 | 14.61 | 24.6 | — | — | — | — | — | |
| OCCAM-MTraining Status=Rely on training, Exemplar Usage=Prior-Free, Multi-Class=true2026.01 | 14.83 | 21.96 | — | — | 0.21 | 2.2 | — | |
| Ours (vanilla)Training=No, Prompt=box2023.06 | 15.67 | 19.44 | — | — | 0.67 | 5.06 | — | |
| Ours (vanilla)Training=No, Prompt=point2023.06 | 15.67 | 19.44 | — | — | 0.67 | 5.06 | — | |
| PFODCi=5, Ca=252019.05 | 16.52 | 21.45 | 80.27 | — | — | — | — | |
| RetinaNetyear=20182019.04 | 16.62 | 22.3 | — | — | — | — | — | |
| RetinaNetMethod category=Detector provided by the benchmark2022.01 | 16.62 | 22.3 | — | — | — | — | — | |
| RetinaNetYear=2017, Method=Detection, Type=Generic2022.08 | 16.62 | 22.3 | — | — | — | — | — | |
| LOCACategory=CAC, Training Dataset=FSC-147, Venue=ICCV'232024.09 | 16.84 | 19.72 | — | — | — | — | — | |
| SAMTraining=No, Prompt=N.A.2023.06 | 16.97 | 20.57 | — | — | 0.7 | 5.3 | — | |
| FamNet+mode=fine-tuned, training_data=CARPK train set2021.04 | 18.19 | 33.66 | — | — | — | — | — | |
| FamNetfine-tuned=true2022.03 | 18.19 | — | — | 33.66 | — | — | — | |
| FamNetMethod category=Few-shot counting method, Evaluation protocol=trained and evaluated by current authors with official code2022.01 | 18.19 | 33.66 | — | — | — | — | — | |
| FamNetFine-Tuned=true, Reference protocol=Reference-based2022.05 | 18.19 | 33.66 | — | — | — | — | — | |
| FamNetYear=2021, Method=Regression, Type=Generic2022.08 | 18.19 | 33.66 | — | — | — | — | — | |
| FamNet+Training=Yes, Prompt=box2023.06 | 18.19 | 33.66 | — | — | — | — | — | |
| CounTRtraining_dataset=FSC-147, supervision=supervised2023.07 | 19.62 | 29.7 | — | — | — | — | 0.57 | |
| VLBaseEvaluation Protocol=zero-shot, Stage=1, Class Name=Checkmark, Train Dataset=FSC1472023.12 | 20.47 | 24.33 | — | — | — | — | — | |
| RCCzero-shot=true2025.02 | 21.38 | 26.61 | — | — | — | — | — | |
| RCCCategory=EFC, Training Dataset=FSC-147, Venue=CVPR'232024.09 | 21.38 | 26.61 | — | — | — | — | — | |
| PFODCi=5, Ca=102019.05 | 22.47 | 27.63 | 73.46 | — | — | — | — | |
| PFODCi=5, Ca=52019.05 | 23.63 | 26.36 | 72.44 | — | — | — | — | |
| LPNCi=989, Ca=ALL2019.05 | 23.8 | 36.79 | — | — | — | — | — | |
| LPN Countingyear=20172019.04 | 23.8 | 36.79 | — | — | — | — | — | |
| Spatially Regularized RPN2021.04 | 23.8 | 36.79 | — | — | — | — | — | |
| LPNMethod category=Single-class car counting method2022.01 | 23.8 | 36.79 | — | — | — | — | — | |
| LPNYear=2017, Method=Detection, Type=Generic2022.08 | 23.8 | 36.79 | — | — | — | — | — | |
| Faster R-CNNyear=20152019.04 | 24.32 | 37.62 | — | — | — | — | — | |
| Faster RCNNvariant=RPN-small2021.04 | 24.32 | 37.62 | — | — | — | — | — | |
| S-RPNMethod category=Detector provided by the benchmark2022.01 | 24.32 | 37.62 | — | — | — | — | — | |
| S-RPNYear=2017, Method=Detection, Type=Generic2022.08 | 24.32 | 37.62 | — | — | — | — | — | |
| GPT-4oSupervision=Zero-shot/MLLM (Z)2026.06 | 24.33 | 36.5 | — | — | — | — | — | |
| UniLIP-3BSupervision=Zero-shot/MLLM (Z)2026.06 | 26.87 | 33.48 | — | — | — | — | — | |
| FamNet-mode=pre-trained, training_data=FSC-147 (excluding car category)2021.04 | 28.84 | 44.47 | — | — | — | — | — | |
| FamNetfine-tuned=false2022.03 | 28.84 | — | — | 44.47 | — | — | — | |
| FamNetFine-Tuned=false, Reference protocol=Reference-based2022.05 | 28.84 | 44.47 | — | — | — | — | — | |
| FamNetSetting=Few-shot2022.11 | 28.84 | 44.47 | — | — | — | — | — | |
| FamNetEvaluation Protocol=few-shot, Stage=1, Class Name=X, Train Dataset=FSC1472023.12 | 28.84 | 44.47 | — | — | — | — | — | |
| FamNetBackBone=ResNet-502024.05 | 28.84 | 44.47 | — | — | — | — | — | |
| Grounding DINOzero-shot=true2025.02 | 29.72 | 31.6 | — | — | — | — | — | |
| UnCounTRbackbone=DINO ViT-B/16, supervision=self-supervised2023.07 | 30.35 | 35.67 | — | — | — | — | 0.54 |