Machine-learning for photoplethysmography analysis: Benchmarking feature, image, and signal-based approaches

Mohammad Moulaeifard; Loic Coquelin; Mantas Rinkevičius; Andrius Sološenko; Oskar Pfeffer; Ciaran Bench; Nando Hegemann; Sara Vardanega; Manasi Nandi; Jordi Alastruey; Christian Heiss; Vaidotas Marozas; Andrew Thompson; Philip J. Aston; Peter H. Charlton; Nils Strodthoff

Machine-learning for photoplethysmography analysis: Benchmarking feature, image, and signal-based approaches

Mohammad Moulaeifard, Loic Coquelin, Mantas Rinkevičius, Andrius Sološenko, Oskar Pfeffer, Ciaran Bench, Nando Hegemann, Sara Vardanega, Manasi Nandi, Jordi Alastruey, Christian Heiss, Vaidotas Marozas, Andrew Thompson, Philip J. Aston, Peter H. Charlton, Nils Strodthoff

Research output: Working paper/Preprint › Preprint

23 Downloads (Pure)

Abstract

Photoplethysmography (PPG) is a widely used non-invasive physiological sensing technique, suitable for various clinical applications. Such clinical applications are increasingly supported by machine learning methods, raising the question of the most appropriate input representation and model choice. Comprehensive comparisons, in particular across different input representations, are scarce. We address this gap in the research landscape by a comprehensive benchmarking study covering three kinds of input representations, interpretable features, image representations and raw waveforms, across prototypical regression and classification use cases: blood pressure and atrial fibrillation prediction. In both cases, the best results are achieved by deep neural networks operating on raw time series as input representations. Within this model class, best results are achieved by modern convolutional neural networks (CNNs). but depending on the task setup, shallow CNNs are often also very competitive. We envision that these results will be insightful for researchers to guide their choice on machine learning tasks for PPG data, even beyond the use cases presented in this work.

Original language	Undefined/Unknown
Publication status	Published - 27 Feb 2025

Keywords

cs.LG
eess.SP

Access to Document

2502.19949v1

Cite this

@techreport{7b3aecef9d254d80b99eeb8baf9bb3c4,

title = "Machine-learning for photoplethysmography analysis: Benchmarking feature, image, and signal-based approaches",

abstract = "Photoplethysmography (PPG) is a widely used non-invasive physiological sensing technique, suitable for various clinical applications. Such clinical applications are increasingly supported by machine learning methods, raising the question of the most appropriate input representation and model choice. Comprehensive comparisons, in particular across different input representations, are scarce. We address this gap in the research landscape by a comprehensive benchmarking study covering three kinds of input representations, interpretable features, image representations and raw waveforms, across prototypical regression and classification use cases: blood pressure and atrial fibrillation prediction. In both cases, the best results are achieved by deep neural networks operating on raw time series as input representations. Within this model class, best results are achieved by modern convolutional neural networks (CNNs). but depending on the task setup, shallow CNNs are often also very competitive. We envision that these results will be insightful for researchers to guide their choice on machine learning tasks for PPG data, even beyond the use cases presented in this work.",

keywords = "cs.LG, eess.SP",

author = "Mohammad Moulaeifard and Loic Coquelin and Mantas Rinkevi{\v c}ius and Andrius Solo{\v s}enko and Oskar Pfeffer and Ciaran Bench and Nando Hegemann and Sara Vardanega and Manasi Nandi and Jordi Alastruey and Christian Heiss and Vaidotas Marozas and Andrew Thompson and Aston, {Philip J.} and Charlton, {Peter H.} and Nils Strodthoff",

note = "39 pages, 9 figures, code available at https://gitlab.com/qumphy/d1-code",

year = "2025",

month = feb,

day = "27",

language = "Undefined/Unknown",

type = "WorkingPaper",

}

TY - UNPB

T1 - Machine-learning for photoplethysmography analysis

T2 - Benchmarking feature, image, and signal-based approaches

AU - Moulaeifard, Mohammad

AU - Coquelin, Loic

AU - Rinkevičius, Mantas

AU - Sološenko, Andrius

AU - Pfeffer, Oskar

AU - Bench, Ciaran

AU - Hegemann, Nando

AU - Vardanega, Sara

AU - Nandi, Manasi

AU - Alastruey, Jordi

AU - Heiss, Christian

AU - Marozas, Vaidotas

AU - Thompson, Andrew

AU - Aston, Philip J.

AU - Charlton, Peter H.

AU - Strodthoff, Nils

N1 - 39 pages, 9 figures, code available at https://gitlab.com/qumphy/d1-code

PY - 2025/2/27

Y1 - 2025/2/27

N2 - Photoplethysmography (PPG) is a widely used non-invasive physiological sensing technique, suitable for various clinical applications. Such clinical applications are increasingly supported by machine learning methods, raising the question of the most appropriate input representation and model choice. Comprehensive comparisons, in particular across different input representations, are scarce. We address this gap in the research landscape by a comprehensive benchmarking study covering three kinds of input representations, interpretable features, image representations and raw waveforms, across prototypical regression and classification use cases: blood pressure and atrial fibrillation prediction. In both cases, the best results are achieved by deep neural networks operating on raw time series as input representations. Within this model class, best results are achieved by modern convolutional neural networks (CNNs). but depending on the task setup, shallow CNNs are often also very competitive. We envision that these results will be insightful for researchers to guide their choice on machine learning tasks for PPG data, even beyond the use cases presented in this work.

AB - Photoplethysmography (PPG) is a widely used non-invasive physiological sensing technique, suitable for various clinical applications. Such clinical applications are increasingly supported by machine learning methods, raising the question of the most appropriate input representation and model choice. Comprehensive comparisons, in particular across different input representations, are scarce. We address this gap in the research landscape by a comprehensive benchmarking study covering three kinds of input representations, interpretable features, image representations and raw waveforms, across prototypical regression and classification use cases: blood pressure and atrial fibrillation prediction. In both cases, the best results are achieved by deep neural networks operating on raw time series as input representations. Within this model class, best results are achieved by modern convolutional neural networks (CNNs). but depending on the task setup, shallow CNNs are often also very competitive. We envision that these results will be insightful for researchers to guide their choice on machine learning tasks for PPG data, even beyond the use cases presented in this work.

KW - cs.LG

KW - eess.SP

M3 - Preprint

BT - Machine-learning for photoplethysmography analysis

ER -