PMID- 34673496
OWN - NLM
STAT- MEDLINE
DCOM- 20220405
LR  - 20220613
IS  - 1940-9990 (Electronic)
IS  - 1932-4545 (Linking)
VI  - 15
IP  - 6
DP  - 2021 Dec
TI  - Q-PPG: Energy-Efficient PPG-Based Heart Rate Monitoring on Wearable Devices.
PG  - 1196-1209
LID - 10.1109/TBCAS.2021.3122017 [doi]
AB  - Hearth Rate (HR) monitoring is increasingly performed in wrist-worn devices using 
      low-cost photoplethysmography (PPG) sensors. However, Motion Artifacts (MAs) 
      caused by movements of the subject's arm affect the performance of PPG-based HR 
      tracking. This is typically addressed coupling the PPG signal with acceleration 
      measurements from an inertial sensor. Unfortunately, most standard approaches of 
      this kind rely on hand-tuned parameters, which impair their generalization 
      capabilities and their applicability to real data in the field. In contrast, 
      methods based on deep learning, despite their better generalization, are 
      considered to be too complex to deploy on wearable devices. In this work, we 
      tackle these limitations, proposing a design space exploration methodology to 
      automatically generate a rich family of deep Temporal Convolutional Networks 
      (TCNs) for HR monitoring, all derived from a single "seed" model. Our flow 
      involves a cascade of two Neural Architecture Search (NAS) tools and a 
      hardware-friendly quantizer, whose combination yields both highly accurate and 
      extremely lightweight models. When tested on the PPG-Dalia dataset, our most 
      accurate model sets a new state-of-the-art in Mean Absolute Error. Furthermore, 
      we deploy our TCNs on an embedded platform featuring a STM32WB55 microcontroller, 
      demonstrating their suitability for real-time execution. Our most accurate 
      quantized network achieves 4.41 Beats Per Minute (BPM) of Mean Absolute Error 
      (MAE), with an energy consumption of 47.65 mJ and a memory footprint of 412 kB. 
      At the same time, the smallest network that obtains a MAE 8 BPM, among those 
      generated by our flow, has a memory footprint of 1.9 kB and consumes just 1.79 mJ 
      per inference.
FAU - Burrello, Alessio
AU  - Burrello A
FAU - Pagliari, Daniele Jahier
AU  - Pagliari DJ
FAU - Risso, Matteo
AU  - Risso M
FAU - Benatti, Simone
AU  - Benatti S
FAU - Macii, Enrico
AU  - Macii E
FAU - Benini, Luca
AU  - Benini L
FAU - Poncino, Massimo
AU  - Poncino M
LA  - eng
PT  - Journal Article
PT  - Research Support, Non-U.S. Gov't
DEP - 20220217
PL  - United States
TA  - IEEE Trans Biomed Circuits Syst
JT  - IEEE transactions on biomedical circuits and systems
JID - 101312520
SB  - IM
MH  - Algorithms
MH  - Artifacts
MH  - Heart Rate/physiology
MH  - *Photoplethysmography
MH  - Signal Processing, Computer-Assisted
MH  - *Wearable Electronic Devices
EDAT- 2021/10/22 06:00
MHDA- 2022/04/05 06:00
CRDT- 2021/10/21 21:06
PHST- 2021/10/22 06:00 [pubmed]
PHST- 2022/04/05 06:00 [medline]
PHST- 2021/10/21 21:06 [entrez]
AID - 10.1109/TBCAS.2021.3122017 [doi]
PST - ppublish
SO  - IEEE Trans Biomed Circuits Syst. 2021 Dec;15(6):1196-1209. doi: 
      10.1109/TBCAS.2021.3122017. Epub 2022 Feb 17.