Video Repetition Counting on UCFRep (test)
11.6MAECountLLM
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| CountLLMEncoder=VideoMAE2025.03 | 11.6 | 83.9 | — | — | |
| CountLLMTraining dataset=RepCount2025.03 | 13.9 | 81.6 | — | — | |
| Si & So+Encoder=R(2+1)D182025.03 | 14.3 | 80 | — | — | |
| Sight & SoundEncoder=R(2+1)D18, Fine-tuning Protocol=Fine-tuned, Modality=multi-modal2024.03 | 14.3 | 80 | — | — | |
| ContextEncoder=RX3D1012025.03 | 14.7 | 79 | — | — | |
| ContextEncoder=RX3D101, Fine-tuning Protocol=Fine-tuned2024.03 | 14.7 | 79 | 2.165 | 0.452 | |
| CountLLMEncoder=VSwinT2025.03 | 15.3 | 80.2 | — | — | |
| ESCountsEncoder=VideoMAE2025.03 | 21.6 | 70.4 | — | — | |
| ESCountsEncoder=ViT-B2024.03 | 21.6 | 70.4 | 1.972 | 0.381 | |
| ESCountsEncoder=RX3D1012024.03 | 24.7 | 73.1 | 2.004 | 0.343 | |
| Levy & WolfEncoder=RX3D1012024.03 | 28.6 | 68 | — | — | |
| ESCountsTraining dataset=RepCount2025.03 | 31.7 | 57.1 | — | — | |
| HTRM-Netvideo encoder=Video Swin Transformer tiny, input resolution=224x224, hidden dimension=512, frames=642024.12 | 34.83 | 51.43 | — | — | |
| MFL2024.12 | 38.79 | 51 | — | — | |
| MFLEncoder=RX3D1012025.03 | 38.8 | 51 | — | — | |
| MFLEncoder=RX3D101, Modality=multi-modal2024.03 | 38.8 | 51 | — | — | |
| TransRAC2024.12 | 44.09 | 43 | — | — | |
| TransRACEncoder=VSwinT2025.03 | 44.1 | 43 | — | — | |
| FullEncoder=VSwinT2025.03 | 46.1 | 33.3 | — | — | |
| Zhang et al.2024.12 | 46.89 | 48 | — | — | |
| SightSound2024.12 | 48.25 | 31.25 | — | — | |
| MFLTraining dataset=RepCount2025.03 | 52.3 | 35 | — | — | |
| TransRACTraining dataset=RepCount2025.03 | 64 | 32.4 | — | — | |
| TransRACEncoder=SwinT2024.03 | 64 | 32.4 | — | — | |
| TransRACtrained_on=RepCount part-A2022.04 | 64.01 | 32.4 | — | — | |
| Context (F)Encoder=RX3D101, Fine-tuning Protocol=Frozen2024.03 | 65.3 | 37.2 | 5.761 | 0.143 | |
| ContextTraining dataset=RepCount2025.03 | 76.2 | 41.2 | — | — | |
| RepNetTraining dataset=RepCount2025.03 | 99.8 | 0.9 | — | — | |
| RepNetEncoder=R2D502025.03 | 99.8 | 0.9 | — | — | |
| RepNetEncoder=R2D502024.03 | 99.8 | — | — | 0.009 | |
| RepNettrained_on=RepCount part-A2022.04 | 99.85 | 0.9 | — | — | |
| RepNet2024.12 | 99.85 | 0.9 | — | — |