Heart Rate Estimation on BUAA
1.13MAEFreqPhys
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| FreqPhysVenue=-2026.04 | 1.13 | 1.31 | 0.99 | |
| SiNCTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=TAS, Quality Criterion=rPPG-VQA2026.04 | 1.82 | 2.13 | 0.99 | |
| Gideon21Training Set=CAS(ME)2 + Celeb-DF (v2), Sampling=TAS, Quality Criterion=rPPG-VQA2026.04 | 1.86 | 2.15 | 1 | |
| Gideon21Training Set=CAS(ME)2 + Celeb-DF (v2), Sampling=WRS, Quality Criterion=rPPG-VQA2026.04 | 1.97 | 2.13 | 1 | |
| FreqPhysVenue=-, Training Sources=P+M → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 2.06 | 2.58 | 0.97 | |
| FreqPhysVenue=-, Training Sources=U+M → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 2.12 | 2.77 | 0.96 | |
| Gideon21Training Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Top/Bottom K, Quality Criterion=SNR2026.04 | 2.13 | 2.78 | 0.98 | |
| FreqPhysVenue=-, Training Sources=Average, Method Category=Deep Learning-based rPPG Methods2026.04 | 2.15 | 2.72 | 0.97 | |
| FreqPhysVenue=-2026.04 | 2.17 | 2.45 | 0.99 | |
| Gideon21Training Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Top/Bottom K, Quality Criterion=rPPG-VQA2026.04 | 2.19 | 2.55 | 0.99 | |
| Contrast-PhysTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=TAS, Quality Criterion=rPPG-VQA2026.04 | 2.19 | 3.99 | 0.96 | |
| FreqPhysVenue=-, Training Sources=P+U → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 2.26 | 2.81 | 0.99 | |
| Gideon21Training Set=CAS(ME)2 + Celeb-DF (v2), Sampling=WRS, Quality Criterion=SNR2026.04 | 2.29 | 2.79 | 0.99 | |
| SiNCTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=WRS, Quality Criterion=rPPG-VQA2026.04 | 2.34 | 5.16 | 0.92 | |
| Contrast-PhysTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=WRS, Quality Criterion=SNR2026.04 | 2.37 | 3.61 | 0.97 | |
| SiNCTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Top/Bottom K, Quality Criterion=rPPG-VQA2026.04 | 2.49 | 5.63 | 0.9 | |
| Contrast-PhysTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=WRS, Quality Criterion=rPPG-VQA2026.04 | 2.5 | 4.18 | 0.96 | |
| PHASE-NetVenue=CVPR’262026.04 | 2.56 | 3.25 | 0.96 | |
| SiNCTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=WRS, Quality Criterion=COVER2026.04 | 2.56 | 5.69 | 0.9 | |
| Gideon21Training Set=CAS(ME)2, Sampling=-, Quality Criterion=-2026.04 | 2.58 | 4.45 | 0.96 | |
| Gideon21Training Set=CAS(ME)2 + Celeb-DF (v2), Sampling=WRS, Quality Criterion=COVER2026.04 | 2.67 | 4.03 | 0.96 | |
| Contrast-PhysTraining Set=CAS(ME)2, Sampling=-, Quality Criterion=-2026.04 | 2.72 | 3.74 | 0.93 | |
| Contrast-PhysTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Top/Bottom K, Quality Criterion=rPPG-VQA2026.04 | 2.81 | 5.64 | 0.92 | |
| Gideon21Training Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Top/Bottom K, Quality Criterion=COVER2026.04 | 2.83 | 4.36 | 0.96 | |
| Gideon21Training Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Top/Bottom K, Quality Criterion=IPR2026.04 | 2.87 | 4.93 | 0.95 | |
| Gideon21Training Set=CAS(ME)2 + Celeb-DF (v2), Sampling=WRS, Quality Criterion=IPR2026.04 | 2.88 | 4.9 | 0.96 | |
| PHASE-NetVenue=CVPR’26, Training Sources=P+U → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 2.91 | 4.23 | 0.92 | |
| EfficientPhysVenue=WACV’23, Training Sources=U+M → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 3 | 5.18 | 0.89 | |
| Contrast-PhysTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=WRS, Quality Criterion=COVER2026.04 | 3.1 | 6.15 | 0.91 | |
| Gideon21Training Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Random, Quality Criterion=-2026.04 | 3.28 | 4.91 | 0.96 | |
| SiNCTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Top/Bottom K, Quality Criterion=SNR2026.04 | 3.32 | 7.7 | 0.82 | |
| Contrast-PhysTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Top/Bottom K, Quality Criterion=SNR2026.04 | 3.33 | 7.3 | 0.87 | |
| PHASE-NetVenue=CVPR’26, Training Sources=Average, Method Category=Deep Learning-based rPPG Methods2026.04 | 3.48 | 5.21 | 0.89 | |
| PHASE-NetVenue=CVPR’26, Training Sources=U+M → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 3.51 | 5.18 | 0.89 | |
| RhythmFormerVenue=PR’25, Training Sources=P+U → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 3.9 | 6.51 | 0.82 | |
| SiNCTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=WRS, Quality Criterion=SNR2026.04 | 3.9 | 8.57 | 0.78 | |
| EfficientPhysVenue=WACV’23, Training Sources=Average, Method Category=Deep Learning-based rPPG Methods2026.04 | 3.92 | 6.79 | 0.79 | |
| PHASE-NetVenue=CVPR’26, Training Sources=P+M → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 4.03 | 6.21 | 0.85 | |
| EfficientPhysVenue=WACV’23, Training Sources=P+M → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 4.15 | 7.14 | 0.77 | |
| RhythmFormerVenue=PR’25, Training Sources=P+M → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 4.32 | 6.7 | 0.82 | |
| Contrast-PhysTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Top/Bottom K, Quality Criterion=COVER2026.04 | 4.45 | 9.39 | 0.79 | |
| SiNCTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Top/Bottom K, Quality Criterion=COVER2026.04 | 4.46 | 11.17 | 0.66 | |
| EfficientPhysVenue=WACV’23, Training Sources=P+U → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 4.6 | 8.06 | 0.72 | |
| RhythmFormerVenue=PR’25, Training Sources=Average, Method Category=Deep Learning-based rPPG Methods2026.04 | 4.81 | 8.15 | 0.71 | |
| Contrast-PhysTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Top/Bottom K, Quality Criterion=IPR2026.04 | 4.85 | 9.2 | 0.82 | |
| POSVenue=TBE’16, Training Sources=P+U → B, Method Category=Traditional rPPG Methods2026.04 | 5.04 | 7.12 | 0.63 | |
| POSVenue=TBE’16, Training Sources=P+M → B, Method Category=Traditional rPPG Methods2026.04 | 5.04 | 7.12 | 0.63 | |
| POSVenue=TBE’16, Training Sources=U+M → B, Method Category=Traditional rPPG Methods2026.04 | 5.04 | 7.12 | 0.63 | |
| POSVenue=TBE’16, Training Sources=Average, Method Category=Traditional rPPG Methods2026.04 | 5.04 | 7.12 | 0.63 | |
| POSVenue=TBE’162026.04 | 5.04 | 7.12 | 0.63 | |
| SiNCTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=WRS, Quality Criterion=IPR2026.04 | 5.72 | 12.67 | 0.59 | |
| SiNCTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Top/Bottom K, Quality Criterion=IPR2026.04 | 5.86 | 13.17 | 0.54 | |
| PHASE-NetVenue=CVPR’262026.04 | 5.89 | 7.89 | 0.48 | |
| Gideon21Training Set=CAS(ME)2 + Celeb-DF (v2), Sampling=-, Quality Criterion=-2026.04 | 5.97 | 10.83 | 0.72 | |
| RhythmFormerVenue=PR’252026.04 | 6.04 | 10.84 | 0.42 | |
| CHROMVenue=TBE’13, Training Sources=P+U → B, Method Category=Traditional rPPG Methods2026.04 | 6.09 | 8.29 | 0.51 | |
| CHROMVenue=TBE’13, Training Sources=P+M → B, Method Category=Traditional rPPG Methods2026.04 | 6.09 | 8.29 | 0.51 | |
| CHROMVenue=TBE’13, Training Sources=U+M → B, Method Category=Traditional rPPG Methods2026.04 | 6.09 | 8.29 | 0.51 | |
| CHROMVenue=TBE’13, Training Sources=Average, Method Category=Traditional rPPG Methods2026.04 | 6.09 | 8.29 | 0.51 | |
| CHROMVenue=TBE’132026.04 | 6.09 | 8.29 | 0.51 | |
| RhythmFormerVenue=PR’25, Training Sources=U+M → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 6.2 | 11.23 | 0.49 | |
| PhysLLMMode=Intra-dataset testing2025.05 | 6.48 | 8.48 | 0.63 | |
| PhysLLMVenue=ICLR’262026.04 | 6.48 | 8.48 | 0.63 | |
| Contrast-PhysTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=WRS, Quality Criterion=IPR2026.04 | 6.52 | 12.13 | 0.67 | |
| GreenMode=Intra-dataset testing2025.05 | 6.89 | 10.39 | 0.6 | |
| GREENVenue=Opt Express’08, Training Sources=P+U → B, Method Category=Traditional rPPG Methods2026.04 | 6.89 | 10.39 | 0.6 | |
| GREENVenue=Opt Express’08, Training Sources=P+M → B, Method Category=Traditional rPPG Methods2026.04 | 6.89 | 10.39 | 0.6 | |
| GREENVenue=Opt Express’08, Training Sources=U+M → B, Method Category=Traditional rPPG Methods2026.04 | 6.89 | 10.39 | 0.6 | |
| GREENVenue=Opt Express’08, Training Sources=Average, Method Category=Traditional rPPG Methods2026.04 | 6.89 | 10.39 | 0.6 | |
| GREENVenue=Opt Express’082026.04 | 6.89 | 10.39 | 0.6 | |
| GREENVenue=Opt Express’082026.04 | 6.89 | 10.39 | 0.6 | |
| Contrast-PhysTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Random, Quality Criterion=-2026.04 | 7.69 | 15.2 | 0.41 | |
| PhysFormerMode=Intra-dataset testing2025.05 | 8.46 | 10.17 | -0.06 | |
| PhysFormerVenue=CVPR’222026.04 | 8.46 | 10.17 | -0.06 | |
| RhythmFormerMode=Intra-dataset testing2025.05 | 9.19 | 11.93 | -0.1 | |
| Contrast-PhysTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=-, Quality Criterion=-2026.04 | 9.38 | 15.89 | 0.52 | |
| PhysFormerVenue=CVPR’22, Training Sources=U+M → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 10.87 | 16.2 | 0.08 | |
| PhysNetMode=Intra-dataset testing2025.05 | 10.89 | 11.7 | -0.04 | |
| PhysNetVenue=BMVC’192026.04 | 10.89 | 11.7 | -0.04 | |
| SiNCTraining Set=CAS(ME)2, Sampling=-, Quality Criterion=-2026.04 | 10.9 | 16.31 | 0.3 | |
| SiNCTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=Random, Quality Criterion=-2026.04 | 10.93 | 17.93 | 0.42 | |
| PhysNetVenue=BMVC’19, Training Sources=U+M → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 11.4 | 16.72 | 0.14 | |
| RhythmMambaVenue=AAAI’252026.04 | 11.99 | 14.22 | 0.16 | |
| PhysNetVenue=BMVC’192026.04 | 12.75 | 16.37 | 0.08 | |
| Gideon21Training Set=Celeb-DF (v2), Sampling=-, Quality Criterion=-2026.04 | 13.6 | 19.07 | 0.13 | |
| SiNCTraining Set=CAS(ME)2 + Celeb-DF (v2), Sampling=-, Quality Criterion=-2026.04 | 13.61 | 19.23 | 0.19 | |
| PhysFormerVenue=CVPR’22, Training Sources=Average, Method Category=Deep Learning-based rPPG Methods2026.04 | 14.65 | 18.88 | 0.06 | |
| PhysFormerVenue=CVPR’22, Training Sources=P+M → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 14.86 | 18.26 | 0.03 | |
| PhysNetVenue=BMVC’19, Training Sources=P+U → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 15.34 | 21.48 | -0.29 | |
| PhysNetVenue=BMVC’19, Training Sources=Average, Method Category=Deep Learning-based rPPG Methods2026.04 | 15.9 | 20.98 | -0.05 | |
| EfficientPhysMode=Intra-dataset testing2025.05 | 16.09 | 16.8 | 0.14 | |
| EfficientPhysVenue=WACV’232026.04 | 16.09 | 16.8 | 0.14 | |
| SiNCTraining Set=Celeb-DF (v2), Sampling=-, Quality Criterion=-2026.04 | 16.23 | 21.27 | 0.11 | |
| PhysFormerVenue=CVPR’22, Training Sources=P+U → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 18.23 | 22.17 | 0.07 | |
| PhysNetVenue=BMVC’19, Training Sources=P+M → B, Method Category=Deep Learning-based rPPG Methods2026.04 | 20.97 | 24.75 | 0.01 | |
| PhysFormerVenue=CVPR’222026.04 | 22.09 | 26.21 | 0.03 | |
| Contrast-PhysTraining Set=Celeb-DF (v2), Sampling=-, Quality Criterion=-2026.04 | 23.24 | 28.55 | -0.08 | |
| EfficientPhysVenue=WACV’232026.04 | 32.3 | 34 | -0.03 |