An Accurate Noninvasive Blood Glucose Measurement System Using Portable Near-Infrared Spectrometer and Transfer Learning Framework

Yu, Yang; Jian-ping, Huang; Zhu, Jigui; Liang, Shili

doi:10.1109/jsen.2020.3025826

Cited by 32 publications

(16 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The predicted value is very close to the chemical measurement value, The calibration model is accurate and reliable.In biomedicine, Weng Weiwei carried out a feasibility study on noninvasive evaluation of muscle hemodynamic response in patients with Duchenne muscular dystrophy by using near-infrared spectroscopy based on DMD [46] . The study shows that this method can measure the muscle hemodynamic response in small muscle vessels affected by no, which is an important factor of muscle ischemia in such patients during exercise.Yu Yong et al developed a portable near-infrared spectrometer blood glucose measurement system based on DMD [47] . As shown in fig.…”

Section: Application Of Dmd Near Infrared Spectrographmentioning

confidence: 99%

Research on near infrared spectroscopy based on DMD

liu

song

2023

Earth and Space: From Infrared to Terahertz (ESIT 2022)

View full text Add to dashboard Cite

Near infrared spectroscopy has been applied to qualitative and quantitative analysis in many fields due to its advantages of fast measurement speed, high accuracy and continuous measurement. Digital micromirror array is an electronic input and output optical mems, semiconductor manufacturing technology combined with the height of the integrated circuit manufacturing technology, make its in resolution, contrast, brightness, grayscale, color fidelity and main performance parameters, such as response time reached the high level, high reliability, fast response speed. In recent years, near-infrared spectroscopy based on wavelength grating and modulation characteristics of DMD has become a research hotspot. This system has the advantages of high signal-to-noise ratio, excellent comprehensive performance and low cost, and has attracted much attention in many application fields. In this paper, the principle, development history, application status and development direction of near infrared spectroscopy based on DMD are reviewed, which provides a basis for the related research of DMD near infrared spectroscopy analysis technology.

show abstract

Section: Application Of Dmd Near Infrared Spectrographmentioning

confidence: 99%

Research on near infrared spectroscopy based on DMD

liu

song

2023

Earth and Space: From Infrared to Terahertz (ESIT 2022)

View full text Add to dashboard Cite

show abstract

“…The feasibility of a cellular automata based tracking method of blood glucose through skin impedance measurement was introduced in [ 54 ]. Recently, Yan et al [ 55 ] using a portable NIR Spectrometer and advanced machine learning models have shown the model based on the combination of synergy interval, genetic algorithm and extreme learning machine as the most accurate for blood glucose detection. A partial least squares method was used in [ 56 ] to obtain a correlation coefficient of 93.2% and a prediction error of 0.23 mmol/L.…”

Section: Optical Spectroscopymentioning

confidence: 99%

A Concise and Systematic Review on Non-Invasive Glucose Monitoring for Potential Diabetes Management

et al. 2022

View full text Add to dashboard Cite

The current standard of diabetes management depends upon the invasive blood pricking techniques. In recent times, the availability of minimally invasive continuous glucose monitoring devices have made some improvements in the life of diabetic patients however it has its own limitations which include painful insertion, excessive cost, discomfort and an active risk due to the presence of a foreign body under the skin. Due to all these factors, the non-invasive glucose monitoring has remain a subject of research for the last two decades and multiple techniques of non-invasive glucose monitoring have been proposed. These proposed techniques have the potential to be evolved into a wearable device for non-invasive diabetes management. This paper reviews research advances and major challenges of such techniques or methods in recent years and broadly classifies them into four types based on their detection principles. These four methods are: optical spectroscopy, photoacoustic spectroscopy, electromagnetic sensing and nanomaterial based sensing. The paper primarily focuses on the evolution of non-invasive technology from bench-top equipment to smart wearable devices for personalized non-invasive continuous glucose monitoring in these four methods. With the rapid evolve of wearable technology, all these four methods of non-invasive blood glucose monitoring independently or in combination of two or more have the potential to become a reality in the near future for efficient, affordable, accurate and pain-free diabetes management.

show abstract

“…This is due to the fact that the bands in NIR spectra are broad, correlated, and highly overlapping which required the presence of mathematical tools to extract the analytical information from these featureless spectra. Nowadays, with the progressive evolution in the NIRS instrument and mathematical resources, NIRS has been successfully explored and widely applied especially in agriculture and recently it has successfully contributed to the post harvested decision support system [5][6][7].…”

Section: Introductionmentioning

confidence: 99%

“…Thus, the ability of TL to utilize knowledge present in labelled training data from a source domain to enhance a model performance in a target domain may be an alternative to address the limitation of calibration transfer. However, the studies of transfer learning for NIRS is limited in numbers; and thus more studies in constructing an efficient model using transfer learning are much needed [6,17]. Thus, this study aims to evaluate and analyse the performance of transferred models from different domain feature spaces (across different harvest seasons) using transfer learning approach.…”

Section: Introductionmentioning

confidence: 99%