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

2025-02-27

AI Summary

This study systematically investigates the performance trade-offs of photoplethysmography (PPG) signal representations and model architectures across two distinct clinical tasks: blood pressure prediction (regression) and atrial fibrillation detection (classification). We conduct the first cross-paradigm benchmark, evaluating three input modalities—handcrafted features, time-frequency images (e.g., STFT and CWT spectrograms), and raw time-series waveforms—paired with both classical machine learning and deep learning models (including CNNs and ResNet variants). Results demonstrate that raw waveform inputs coupled with modern CNN architectures achieve superior and robust performance; even shallow CNNs exhibit strong competitiveness. In contrast, image-based representations yield no significant gains, and handcrafted features suffer from limited generalizability. To foster reproducibility and standardization, we introduce the first open-source, end-to-end, multi-task PPG benchmark framework. This work provides both methodological guidance and an empirical foundation for non-invasive, PPG-driven physiological monitoring.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

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.

Problem

• Benchmarking machine learning for PPG analysis
• Comparing feature, image, and signal-based approaches
• Optimizing input representations for clinical predictions

Innovation

• Deep neural networks for PPG analysis
• Convolutional neural networks on raw data
• Benchmarking feature, image, signal approaches