# Remote Photoplethysmography (rPPG) on Mobile/ARM Hardware Efficiency

Source: [Remote Photoplethysmography (rPPG) on Mobile/ARM Hardware Efficiency](<https://www.sota2.com/research/sota/remote-photoplethysmography-rppg-on-mobile-arm-hardware-efficiency>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:45:32.564Z

Evaluation of the inference speed (latency) for Remote Photoplethysmography (rPPG) models when deployed on mobile or ARM hardware, measured by the time required to process a single video frame.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Remote Photoplethysmography (rPPG) |
| Dataset (including any stated split) | Mobile/ARM Hardware Efficiency |
| Metric | Frame Latency (ms) |
| Unit | TIME |
| Direction | Lower is better |
| Results loaded | 13 of 13 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Frame Latency (ms)](<https://www.sota2.com/research/sota/remote-photoplethysmography-rppg-on-mobile-arm-hardware-efficiency.md?metric=Frame%20Latency%20(ms)>): Lower is better; unit: TIME.

- [Memory Usage (MB)](<https://www.sota2.com/research/sota/remote-photoplethysmography-rppg-on-mobile-arm-hardware-efficiency.md?metric=Memory%20Usage%20(MB)>): Lower is better; unit: SIZE.

- [FLOPs (G)](<https://www.sota2.com/research/sota/remote-photoplethysmography-rppg-on-mobile-arm-hardware-efficiency.md?metric=FLOPs%20(G)>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Frame Latency (ms) | Result date | Report paper |
| --- | --- | --- | --- |
| FacePhys | 9.46 | Not recorded | 6937ad266cfd488761a4e97e |
| TSCAN | 55.2 | Not recorded | 6937ad266cfd488761a4e97e |
| PhysNet | 62.1 | Not recorded | 6937ad266cfd488761a4e97e |
| EFFPhys | 371 | Not recorded | 6937ad266cfd488761a4e97e |
| Beam AI SDK | 500 | Not recorded | 68fa6d921e209131cea3abbe |
| RhyMamba | 889 | Not recorded | 6937ad266cfd488761a4e97e |
| PhysFormer | 2330 | Not recorded | 6937ad266cfd488761a4e97e |
| POS | 2700 | Not recorded | 68fa6d921e209131cea3abbe |
| CHROM | 2800 | Not recorded | 68fa6d921e209131cea3abbe |
| ICA | 3100 | Not recorded | 68fa6d921e209131cea3abbe |
| EfficientPhys-C | 4000 | Not recorded | 68fa6d921e209131cea3abbe |
| TS-CAN | 6000 | Not recorded | 68fa6d921e209131cea3abbe |
| EfficientPhys-T1 | 30000 | Not recorded | 68fa6d921e209131cea3abbe |

### FacePhys

Frame Latency (ms): 9.46. Result ID: 6998847e5698665ec4e5cb72.

[Reporting paper](<https://www.sota2.com/research/paper/6937ad266cfd488761a4e97e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6937ad266cfd488761a4e97e.md>)

Recorded evaluation configuration: {"Params (K)":719}. This record does not assert independent reproduction.

### TSCAN

Frame Latency (ms): 55.2. Result ID: 6998847e5698665ec4e5cb6e.

[Reporting paper](<https://www.sota2.com/research/paper/6937ad266cfd488761a4e97e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69751e209131cea22375.md>)

[Implementation](<https://github.com/xin71/MTTS-CAN>)

Recorded evaluation configuration: {"Params (K)":533}. This record does not assert independent reproduction.

### PhysNet

Frame Latency (ms): 62.1. Result ID: 6998847e5698665ec4e5cb6f.

[Reporting paper](<https://www.sota2.com/research/paper/6937ad266cfd488761a4e97e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa661f1e209131cea0e7d2.md>)

Recorded evaluation configuration: {"Params (K)":770}. This record does not assert independent reproduction.

### EFFPhys

Frame Latency (ms): 371. Result ID: 6998847e5698665ec4e5cb6d.

[Reporting paper](<https://www.sota2.com/research/paper/6937ad266cfd488761a4e97e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e411e209131cea3ebf3.md>)

Recorded evaluation configuration: {"Params (K)":2189}. This record does not assert independent reproduction.

### Beam AI SDK

Frame Latency (ms): 500. Result ID: 698faaeb3a6c314849faf14d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d921e209131cea3abbe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d921e209131cea3abbe.md>)

[Implementation](<https://www.sota2.com/github.com/beamai/BeamAISDK-iOS>)

Recorded evaluation configuration: {"Device":"iPhone 13"}. This record does not assert independent reproduction.

### RhyMamba

Frame Latency (ms): 889. Result ID: 6998847e5698665ec4e5cb71.

[Reporting paper](<https://www.sota2.com/research/paper/6937ad266cfd488761a4e97e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa771e1e209131cea7131e.md>)

Recorded evaluation configuration: {"Params (K)":4936}. This record does not assert independent reproduction.

### PhysFormer

Frame Latency (ms): 2330. Result ID: 6998847e5698665ec4e5cb70.

[Reporting paper](<https://www.sota2.com/research/paper/6937ad266cfd488761a4e97e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c961e209131cea34adb.md>)

Recorded evaluation configuration: {"Params (K)":7395}. This record does not assert independent reproduction.

### POS

Frame Latency (ms): 2700. Result ID: 698faaeb3a6c314849faf151.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d921e209131cea3abbe.md>)

Recorded evaluation configuration: {"Device":"ARM CPU"}. This record does not assert independent reproduction.

### CHROM

Frame Latency (ms): 2800. Result ID: 698faaeb3a6c314849faf152.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d921e209131cea3abbe.md>)

Recorded evaluation configuration: {"Device":"ARM CPU"}. This record does not assert independent reproduction.

### ICA

Frame Latency (ms): 3100. Result ID: 698faaeb3a6c314849faf153.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d921e209131cea3abbe.md>)

Recorded evaluation configuration: {"Device":"ARM CPU"}. This record does not assert independent reproduction.

### EfficientPhys-C

Frame Latency (ms): 4000. Result ID: 698faaeb3a6c314849faf150.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d921e209131cea3abbe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e411e209131cea3ebf3.md>)

Recorded evaluation configuration: {"Device":"ARM CPU"}. This record does not assert independent reproduction.

### TS-CAN

Frame Latency (ms): 6000. Result ID: 698faaeb3a6c314849faf14f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d921e209131cea3abbe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69751e209131cea22375.md>)

[Implementation](<https://github.com/xin71/MTTS-CAN>)

Recorded evaluation configuration: {"Device":"ARM CPU"}. This record does not assert independent reproduction.

### EfficientPhys-T1

Frame Latency (ms): 30000. Result ID: 698faaeb3a6c314849faf14e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d921e209131cea3abbe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e411e209131cea3ebf3.md>)

Recorded evaluation configuration: {"Device":"ARM CPU"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
