You can check the ＂blood routine＂ App by taking a photo on your mobile phone.

Original changguangsuo Light center China optics

Cover image Source: Purdue University

Write | R

01

guide reading

Recently, Sang Mok Park and Michelle A. Visbal-Onufrak of Wilton College of Biomedical Engineering of Purdue University in the United States released a technology in Optica. They have successfully developed an App (mHematology), which can only rely on the smartphone camera to determine the hemoglobin level. The most common clinical laboratory tests can be carried out without blood drawing, which helps to reduce the need for patients to go to outpatient clinics in person, make it easier to monitor critically ill patients, and improve the medical level of low-income and middle-income countries with limited testing laboratories.

02

background introduction

As one of the most common clinical laboratory tests, hemoglobin (Hgb) test is usually used as a preliminary screening for erythropoiesis, so as to check the general health before other specific tests. Spectral analysis is usually used to measure the hemoglobin content of blood because it has a unique light absorption spectrum or fingerprint in the visible wavelength range. However, this type of analysis usually requires huge and expensive optical components, so it is not friendly for patients to deploy this health detection scheme.

The researchers of this project use software to convert the built-in camera of smart phone into hyperspectral imager. Combined with statistical learning technology, this imager can reliably measure hemoglobin level without any hardware modification or accessories, and its mobile application is named MHematography. In addition, the researchers chose the inner eyelid as the sensing site because the microvascular system there is easy to see; It is easy to access and has a relatively uniform red color. The inner eyelid is not affected by skin color, so no personal calibration is needed.

Fig. 1. Spectral super-resolution (SSR) spectroscopy for hemoglobin analysis in mobile health (mHealth).

Image source: vol.7, No.6/June 2020/optica (fig.1).

03

innovation research

The process of hemoglobin measurement using this technology: the patient pulls down the inner eyelid to expose the small blood vessels below. Then, professionals use the smartphone application mHematology to take photos of eyelids (the flash is not aimed directly at the eyeball). Then the spectral super-resolution (SSR) algorithm is used to extract detailed spectral information from RGB photos, and then another algorithm quantifies the hemoglobin content of blood by detecting unique spectral characteristics. Although spectral information is currently extracted by an algorithm on a separate computer, researchers expect that the algorithm can be integrated into mobile applications. In addition, the analysis of multiple linear regression shows that only using RGB data (without the intermediate step of extracting spectral information) may not provide enough information to reliably evaluate the Hgb level in blood.

Figure 2. High-quality spectrum obtained by image-guided hyperspectral scanning system and mHematology mobile application.

Image source: vol.7, No.6/June 2020/optica (fig.2).

3.1

Spectral super-resolution (SSR) algorithm

Hyperspectral data in the visible range is mathematically reconstructed from an RGB image (that is, three-color information from R, G and B channels). In other words, it aims to solve the ill-posed problem as an inverse mapping from down-sampled space (three colors) to dense space (multiple wavelengths).

Fig. 3. Comparison between original spectrum and SSR reconstructed spectrum.

Image source: vol.7, No.6/June 2020/optica (fig.4).

3.2

Quantification of hemoglobin (Hgb) in blood by spectral characteristics

When using a high-resolution spectrum with a large number of wavelengths, we can identify the key potential variables from the spectral intensity, which are responsible for capturing the largest change in the result variable (that is, the blood hemoglobin value). Therefore, this project uses partial least squares regression (PLSR) to transform high-dimensional measured variables into a reduced potential variable space in a way similar to principal component analysis. In order to avoid the possible over-fitting of the prediction model, we determined the optimal number of principal components and verified the spectral blood Hgb prediction model with an independent verification data set.

Figure 4. Performance of SSR blood Hgb measurement using mHematology mobile application.

Image source: vol.7, No.6/June 2020/optica (fig.5).

Clinical tests show that the mobile health test can provide the measurement value equivalent to the traditional blood test, and the measurement range covers a wide range of hemoglobin values. This work shows that data-driven and data-centered light-based research can provide new methods to minimize hardware complexity and promote mobile health. Kim, the leader of the research team, said, "Combining the built-in sensors in today’s smart phones with data-centric methods can speed up innovation and research transformation in this field."

Article information:

The related results were published in Optica with the title of "mhealth spectroscopy of blood hemoglobin with spectral super-resolution".

Paper address:

https://doi.org/10.1364/OPTICA.390409

▶【】

WeChat official account’s timeline has been revised, and many readers feedback that they haven’t seen the updated articles. According to the latest rules, it is suggested that you should visit China Optics WeChat official account several times, read 3-5 articles, and become a "frequent reader" to receive them in time.

Welcome the research group to submit the publicity draft of the results.

Reprinted/cooperated/contributed by the research group, please add WeChat: 447882024.

Banner area

Previous recommendation

Walk into the new issue

Open manuscript: Light:Advanced Manufacturing

ISSN 2689-9620

Periodical website: www.light-am.com

Stay tuned.

New issue: eLight

ISSN 2662-8643

That is, put the weight on the line.

This is the 1329th article published by China Optics. If you think it is helpful, forwarding it to the circle of friends is our greatest recognition.

[Changchun Institute of Optics and Mechanics ·Light Academic Publishing Center-Journal Navigation]

Original title: "App that can check" blood routine "by taking a photo with your mobile phone"

Read the original text