Spectral Transformation-Based PPG Representation for Improved Blood Pressure Estimation Using Transformer Models

DOI

10.2991/978-94-6239-756-9_21How to use a DOI?

Keywords

Acute Myeloid Leukemia; Extracellular vesicles; Microvesicles; Exosomes; Mesenchymal stromal cells

Abstract

The measurement of blood pressure (BP) accurately and non-invasively is still a very important challenge in monitoring cardiovascular health. This exploratory study evaluates multiple spectral conversion methods, Short-Time Fourier Transform (STFT), Continuous Wavelet Transform (CWT), Fast Fourier Transform (FFT), and Cepstrum, for deep learning BP prediction using photoplethysmogram (PPG) and electrocardiogram (ECG) signals. This study aims to address the need for a systematic comparison of spectral conversion methods towards facilitating the development of cuffless blood pressure estimation methods that can support early diagnosis and continuous health monitoring in clinical and remote settings. STFT provided the best trade-off between predictive performance and computational cost. For our final ensemble model we deployed (an ensemble based upon a weighted geometric mean of the six best performing STFT based transformer configurations), the final ensemble achieved MAE values of 3.91 mmHg (DBP) and 4.97 mmHg (SBP), with standard deviations of 6.96 mmHg and 7.88 mmHg, respectively. Pearson correlation coefficients were 0.9220 (DBP) and 0.9391 (SBP). Both STFT models in terms of MAE and their respective standard deviations satisfy the clinical acceptability criteria (as defined by the Association for the Advancement of Medical Instrumentation (AAMI). Overall, these results indicate strong potential for reliable and clinically acceptable cuffless BP estimation using STFT-based features with transformer architectures.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Shyamala Subramanian
AU  - Sashikala Mishra
AU  - Ranjeet Bidwe
PY  - 2026
DA  - 2026/08/31
TI  - Spectral Transformation-Based PPG Representation for Improved Blood Pressure Estimation Using Transformer Models
BT  - Proceedings of the Conference on Bioengineering for Global Health (SYMRESEARCH 2.0 2025)
PB  - Atlantis Press
SP  - 295
EP  - 313
SN  - 2468-5747
UR  - https://doi.org/10.2991/978-94-6239-756-9_21
DO  - 10.2991/978-94-6239-756-9_21
ID  - Subramanian2026
ER  -