Non-invasive non-contact system and method for measuring diabetes mellitus condition using thermal imaging

12023,130

17/291213

October 02, 2020

System and method for measuring diabetes mellitus condition of a subject is disclosed. The disclosed system and method includes thermal sensors for capturing thermal images and/or videos of a body part; and a processing engine to detect a predefined region of the body part in each frame of the captured images and/or videos. The processing engine segments one or more portions from the detected predefined region in each frame of the captured images and/or videos to identify a region of interest comprising major arteries in the segmented portions. Based on the ROI, the engine extracts pixel values, representing biosignals, from each frame of the captured images and/or videos so as to determine one or more parameters associated with the hemodynamic factors and a rate of atherosclerosis of the subject. Further, a risk score for the diabetes mellitus condition based on the determined parameters using computational models is measured.

Aarca Research Inc. (Orange, CT, US)

Aarca Research, Inc. (Orange, CT, US)

1. A system for measuring diabetes mellitus condition of a subject, the system comprising: a set of thermal sensors for capturing any or a combination of one or more thermal images and videos of at least one body part of the subject; and a processing engine operatively coupled to the set of thermal sensors, and comprising one or more processors coupled to a memory, the memory storing a set of instructions executable by the one or more processors to: receive a set of data packets associated with the captured any or a combination of one or more images and videos from the set of thermal sensors; detect a predefined region of the at least one body part of the subject in each frame of the captured any or a combination of the one or more images and videos on receipt of the set of data packets; segment one or more portions from the detected predefined region in each frame of the captured any or a combination of the one or more images and videos; identify a region of interest comprising one or more arteries in the one or more segmented portions in each frame of the captured any or a combination of the one or more images and videos; extract one or more pixel values, representing a set of biosignals, from each frame of the captured any or a combination of the one or more images and videos based on the identified region of interest; determine one or more parameters associated with hemodynamic factors and a rate of atherosclerosis of the subject based on the extracted one or more pixel values representing the set of biosignals; and measure a risk score for the diabetes mellitus condition based on the determined one or more parameters, wherein measuring the risk score is performed using computational models, and wherein for generating the computational models, the processing engine is configured to: calculate signal parameters for one or more signals associated with the one or more arteries of both a healthy subject and a diabetes subject; identify a set of parameters that correlate to complications associated with the diabetes condition using a principal component analysis on the calculated signal parameters; determine patterns and differences among the parameters between diabetes and healthy subjects by using statistical methods and visualization; and train the computational models by using machine learning units comprising clustering models, logistic regression, random forest, or neural network models on the parameters.

2. The system as claimed in claim 1 , wherein the set of thermal sensors are selected from the group consisting of a digital camera, a digital single-lens reflex (DSLR) camera, an infrared camera, and a thermal camera, and wherein the set of thermal sensors sense heat radiation or infrared radiation emitted from the body part of the subject and renders the one or more images and videos representing a spatial intensity of the heat radiation or the infrared radiation.

3. The system as claimed in claim 1 , wherein the determined one or more parameters are associated with potential biomarkers of hemodynamic imbalances and atherosclerosis, and wherein measurement of the risk score for the diabetes mellitus condition is based on comparison of the determined one or more parameters with predetermined set of reference parameters that are stored in a database operatively coupled to the processing engine.

4. The system as claimed in claim 1 , wherein the subject is a human.

5. The system as claimed in claim 1 , wherein the at least one body parts of the subject is an anterior face of the subject, and wherein the segment one or more portions are associated with a forehead of the subject.

6. The system as claimed in claim 5 , wherein the identified region of interest is associated with a forehead region of the subject comprises frontal branches of the arteries.

7. The system as claimed in claim 6 , wherein the one or more processors are configured to segment the identified region of interest from each of the captured any or a combination of the one or more images and videos based on a difference between thermal intensity along frontal branches of the arteries and thermal intensity in other regions of the forehead.

8. The system as claimed in claim 7 , wherein the identified region of interest is segmented using any or a combination of morphological operations, otsu thresholding, edge detection and contour approximations techniques.

9. The system as claimed in claim 1 , wherein the one or more processors are configured to execute a first set of instructions associated with image filtering and enhancing techniques on each of the captured any or a combination of the one or more images and videos for removing noise and improving quality.

10. The system as claimed in claim 1 , wherein the one or more processors are configured to execute a second set of instruction associated with image processing including face detection and landmark detection to detect the predefined region of the at least one body part in each frame of the captured any or a combination of the one or more images and videos.

11. The system as claimed in claim 1 , wherein the one or more processors are configured to perform spatial transformation on the identified region of interest to obtain a quantitative representation of a pattern observed in each frame of the captured any or a combination of the one or more images and videos, representing a set of biosignals waveforms along the one or more arteries associated with pulsatile nature of blood flow.

12. The system as claimed in claim 1 , wherein the one or more processors are configured to normalize and filter the one or more extracted pixel values representing the set of biosignals, and wherein the one or more processors are configured to determine time domain values by applying statistical analysis on the filtered pixel values.

13. The system as claimed in claim 12 , wherein the one or more processors are configured to determine frequency domain values by applying Fast Fourier Transform and frequency filtering technique on the determined time domain values.

14. The system as claimed in claim 13 , wherein the one or more processors are configured to determine, using signal processing techniques, signal parameters comprising time and frequency domain parameters based on the determined frequency domain values and time domain values.

15. The system as claimed in claim 14 , wherein the determined time and frequency domain parameters comprises any or a combination of average intensity, signal amplitude, signal period, signal entropy, signal power spectral density, histogram and peak count, and wherein the time and frequency domain parameters are also associated with any or a combination of the hemodynamics factors, rate of atherosclerosis, general healthiness of the artery itself or physiological data indicating core temperature, blood flow velocity, blood density, arterial stiffness, and oxygen saturation in blood.

16. The system as claimed in claim 1 , wherein the measurement of the risk score for the diabetes mellitus condition of the subject considers demographics and medical history of the subject along with the determined one or more parameters.

17. The system as claimed in claim 1 , wherein the determined one or more parameters are associated with a degree of hyperglycemia of the subject.

18. A method for measuring diabetes mellitus condition of a subject, the method comprising: capturing, by a set of thermal sensors, any or a combination of one or more thermal images and videos of at least one body part of the subject; receiving, by a processing engine, a set of data packets associated with the captured any or a combination of one or more images and videos from the set of thermal sensors operatively coupled to the processing engine; detecting, by the processing engine, a predefined region of the at least one body part of the subject in each frame of the captured any or a combination of the one or more images and videos on receipt of the set of data packets; segmenting, by the processing engine, one or more portions from the detected predefined region in each frame of the captured any or a combination of the one or more images and videos; identifying, by the processing engine, a region of interest comprising one or more arteries in the one or more segmented portions in each frame of the captured any or a combination of the one or more images and videos; extracting, by the processing engine, one or more pixel values, representing a set of biosignals, from each frame of the captured any or a combination of the one or more images and videos based on the identified region of interest; determining, by the processing engine, one or more parameters associated with hemodynamic factors and a rate of atherosclerosis of the subject based on the extracted one or more pixel values representing the set of biosignals; and measuring, by the processing engine, a risk score for the diabetes mellitus condition based on the determined one or more parameters, wherein measuring the risk score is performed using computational models, and wherein generation of the computational models comprises: calculating, by the processing engine, signal parameters for one or more signals associated with the one or more arteries in both a healthy subject and in a diabetes subject; identifying, by the processing engine, a set of parameters that correlate to complications associated with the diabetes condition using a principal component analysis on the calculated signal parameters; determining, by the processing engine, patterns and differences among the parameters between diabetes and healthy subjects by using statistical methods and visualization; and training, by the processing engine, the computational models by using machine learning units comprising clustering models, logistic regression, random forest, or neural network models on the parameters.

20100172567 July 2010 Prokoski
20160180050 June 2016 Holmes
20170079533 March 2017 Robinson
20170238805 August 2017 Addison
20170238842 August 2017 Jacquel
20200337573 October 2020 Fukuda

Sivanandam, S., et al. “Medical thermography: a diagnostic approach for type 2 diabetes based on non-contact infrared thermal imaging.” Endocrine 42 (2012): 343-351 (Year: 2012). cited by examiner
Bandyopadhyay, Asok, Amit Chaudhuri, and Himanka Sekhar Mondal. “IR based intelligent image processing techniques for medical applications.” 2016 SAI Computing Conference (SAI). IEEE, 2016. (Year: 2016). cited by examiner
Gault, Travis R., et al. “Extraction of the superficial facial vasculature, vital signs waveforms and rates using thermal imaging.” 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition-Workshops. IEEE, 2010. (Year: 2010). cited by examiner
Zhang, Dongfeng, Weijing Wang, and Fang Li. “Association between resting heart rate and coronary artery disease, stroke, sudden death and noncardiovascular diseases: a meta-analysis.” CMAJ 188.15 (2016): E384-E392 (Year: 2016). cited by examiner
Kamadi, Vsrp Varma, Appa Rao Allam, and Sita Mahalakshmi Thummala. “A computational intelligence technique for the effective diagnosis of diabetic patients using principal component analysis (PCA) and modified fuzzy SLIQ decision tree approach.” Applied Soft Computing 49 (2016): 137-145. (Year: 2016). cited by examiner

edspgr.12023130