An Hb Concentration Measurement Method Based on Multimodal Information Fusion and APSO-SVR
Accurate measurement of hemoglobin (Hb) concentration during hemodialysis is essential for monitoring hemoconcentration and evaluating dialysis effectiveness. Existing optical Hb monitoring methods predominantly rely on single-wavelength signals, which are insufficient to fully characterize the complex optical properties of whole blood. Moreover, their measurement accuracy is susceptible to variations in blood temperature, blood pump flow rate, and scattering characteristics. To address these limitations, this study proposes an Hb concentration measurement method based on multimodal information fusion and adaptive particle swarm optimization–support vector regression (APSO-SVR). To investigate the effects of blood temperature, blood pump flow rate, and scattering variations on optical Hb measurement, the rheological properties and optical responses of blood under different conditions were analyzed, and a dual-wavelength measurement scheme was developed based on the distinct optical responses of blood at 660 nm and 940 nm. Subsequently, dual-wavelength optical features and temperature–flow coupling features were constructed using calibrated photoelectric signals, temperature measurements, and blood pump flow rate information. Pearson correlation coefficient (PCC) analysis and minimum redundancy maximum relevance (mRMR) were applied for feature selection, followed by optimization of the SVR parameters using APSO. The proposed method was then evaluated through experiments involving different combinations of temperature and flow rate conditions and comparisons with other prediction models. The results showed that the APSO-SVR model with selected dual-wavelength features achieved good Hb concentration prediction performance, with a root mean square error (RMSE) of 4.34 g/L and a coefficient of determination (R2) of 0.871. The proposed method improved Hb concentration prediction accuracy under different stabilized experimental conditions, demonstrating its potential for optical Hb monitoring. Further validation under dynamic conditions is required to assess its applicability to continuous monitoring during hemodialysis.
Authors
- Wang Fang (ORCID: https://orcid.org/0000-0003-3518-2548)
- Laihu Peng (ORCID: https://orcid.org/0000-0002-5932-104X)
- Zhuosheng Sun (ORCID: https://orcid.org/0009-0006-0855-259X)
- Xiaodong Wang
- Jianqiang Li
- Zexin Zhang
Institutions
- Zhejiang Sci-Tech University (CN)
Publication Details
- Journal
- Sensors
- Published
- 2026-11-20
- DOI
- https://doi.org/10.3390/s26227342
- Primary Topic
- Optical Imaging and Spectroscopy Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Key Research and Development Program of China