Paperdoi:10.1109/JIOT.2026.3716950
MSCFMamba: Lightweight Mamba Model With Transfer Learning for Continuous Cuffless Blood Pressure Monitoring Using PPG and BCG
IEEE Internet of Things Journal (2026; S2 publication date 2026-10-15)
TL;DR
IoMT smart-chair continuous cuffless BP from synchronized PPG + BCG. Fusion-multimodal Markov transition field (FMTF) encodes the pair into a 2-D map; a lightweight multiscale CNN–Mamba (MSCFMamba, 1.71 M params) regresses BP. Transfer learning: pretrain on public Kansas State University (KS) set, fine-tune on a private chair set.
- Form factor: chair
Abstract
Continuous blood pressure (BP) monitoring is essential for hypertension management, yet conventional cuff-based methods are unsuitable for long-term use. This study proposes an Internet of Medical Things (IoMT)-based smart chair (SC) framework for continuous cuffless BP monitoring using synchronized photoplethysmography (PPG) and ballistocardiography (BCG) signals. To preserve complementary multimodal information while reducing model complexity, a fusion-multimodal Markov transition field (FMTF) method is introduced to encode PPG-BCG dynamics into a compact unified 2-D representation, which is then processed by a lightweight multiscale CNN fusion Mamba (MSCFMamba) model. To improve cross-domain adaptation, transfer learning is performed by pretraining on the public Kansas State University (KS) dataset and fine-tuning on a private SC dataset. The proposed model achieves mean absolute error (MAE) ± std of 4.52 ± 1.48 mmHg and 3.08 ± 0.68 mmHg for systolic and diastolic BP, respectively, with only 1.71 M parameters, while meeting Association for the Advancement of Medical Instrumentation (AAMI)/British Hypertension Society (BHS) standards. In addition, clinically acceptable performance is maintained with only 10% subject-specific calibration data. These results demonstrate that the proposed IoMT-based SC provides a practical and compact solution for continuous home-based cuffless BP monitoring in cloud-connected IoMT environments with resource-constrained sensing devices.
Verbatim from Semantic Scholar (opens in a new tab) · Abstract via Semantic Scholar; © the authors/publisher
Results on ChatPPG leaderboards
All leaderboardsNo results from this paper on ChatPPG leaderboards yet. Browse the leaderboards.
Tasks & signals
Tasks
Signals
Keywords
How it works
Cite this paper
APA
Anh, L. T., Le, T. D., Vo, T., Le, Q. P., Nguyen, T. M., Doan, V. H. M., Ha, D., Ahn, J. S., Choi, J., & Oh, J. (2026). MSCFMamba: Lightweight Mamba Model With Transfer Learning for Continuous Cuffless Blood Pressure Monitoring Using PPG and BCG. IEEE Internet of Things Journal. https://doi.org/10.1109/JIOT.2026.3716950
BibTeX
@article{anh2026mscfmamba,
title = {{MSCFMamba: Lightweight Mamba Model With Transfer Learning for Continuous Cuffless Blood Pressure Monitoring Using PPG and BCG}},
author = {Anh, Le The and Le, Thanh Dat and Vo, T. and Le, Quy Phuong and Nguyen, T. M. and Doan, Vu Hoang Minh and Ha, Dogeon and Ahn, Jae Sung and Choi, Jaeyeop and Oh, Junghwan},
year = {2026},
journal = {IEEE Internet of Things Journal},
doi = {10.1109/JIOT.2026.3716950}
}Built from the hub record. Check it against the publisher before submitting.
Added to the ChatPPG research hub
