EP-4740843-A1 - SYSTEM AND METHOD FOR UNOBTRUSIVE OEDEMA SEVERITY CLASSIFICATION AND QUANTIFICATION USING MICROWAVE SENSING

Abstract

Monitoring and detection of oedema severity is crucial for effective treatment of the underlying ailments. Conventional methods involve setups that require harmful radiation, are expensive, tedious, and bulky. The present disclosure provides a system and method for unobtrusive oedema severity classification and quantification using microwave sensing. A vector network analyzer, a microstrip RF patch antenna customized for operation at 4 GHz under dielectric loading of a numerical human phantom and associated RF components are used for oedema severity monitoring of a sample under test. For oedema severity classification, a machine learning (ML) based approach is implemented making use of reflection parameters, and accordingly physics-based electromagnetic features are extracted. The physics-based electromagnetic features are used to classify the sample under test into a specific oedema severity category. Further, a water percentage is quantified using regression models for the sample under test based on the oedema severity category.

Inventors

• KHASNOBISH, Anwesha
• Mazumder, Annesha
• BHARADWAJ, VEDULA KIRAN
• CHAKRAVARTY, TAPAS
• AKHTAR, Mohammad Jaleel

Assignees

• Tata Consultancy Services Limited

Dates

Publication Date

20260513

Application Date

20250922

Claims (18)

1. A processor implemented method (200), comprising: acquiring (202), via one or more hardware processors, a plurality of scan measurements at a plurality of locations of a sample under test using a tuned antenna, wherein the tuned antenna is a microstrip patch antenna operating at a specific operating frequency with one or more optimized dimensions and positioned at a predefined distance from a numerical biological phantom; determining (204), via the one or more hardware processors, a plurality of reflection parameters from the plurality of scan measurements at each location from the plurality of locations of the sample under test using a vector network analyzer, wherein the vector network analyzer utilizes microwave sensing for determining the plurality of reflection parameters; extracting (206), via the one or more hardware processors, a plurality of physics-based electromagnetic (EM) features using a predetermined resonant frequency and a corresponding value of the plurality of reflection parameters at the predetermined resonant frequency to obtain a feature vector, wherein the feature vector comprises a training feature vector and a testing feature vector; classifying (208), via the one or more hardware processors, the sample under test into one of (i) a first oedema severity category, and (ii) a second oedema severity category, using a trained classifier model for the testing feature vector, wherein the trained classifier model is trained using the training feature vector; and quantifying (210), via the one or more hardware processors, the oedema based on a water percentage in the sample under test that is classified into the first oedema severity category and the second oedema severity category, using a trained regression model for a specific set of testing feature vector from the testing feature vector.
2. The processor implemented method as claimed in claim 1, wherein the sample under test is a biological tissue.
3. The processor implemented method as claimed in claim 1, wherein the specific operating frequency of the tuned antenna is 4 Giga Hertz (GHz) and the predefined distance of the tuned antenna from the numerical human phantom is 10 millimeter (mm).
4. The processor implemented method as claimed in claim 1, wherein the plurality of physics-based electromagnetic (EM) features specify water accumulation in the sample under test that represents a detected oedema.
5. The processor implemented method as claimed in claim 1, wherein the first oedema severity category represents a low oedema severity category, and the second oedema severity category represents a medium oedema severity category.
6. The processor implemented method as claimed in claim 1, wherein the step of classifying comprises classifying the sample under test into a third oedema severity category that represents a low oedema severity category.
7. A system (100), comprising: a memory (102) storing instructions; one or more communication interfaces (106); and one or more hardware processors (104) coupled to the memory (102) via the one or more communication interfaces (106), wherein the one or more hardware processors (104) are configured by the instructions to: acquire a plurality of scan measurements at a plurality of locations of a sample under test using a tuned antenna, wherein the tuned antenna is a microstrip patch antenna operating at a specific operating frequency with one or more optimized dimensions and positioned at a predefined distance from a numerical biological phantom; determine a plurality of reflection parameters from the plurality of scan measurements at each location from the plurality of locations of the sample under test using a vector network analyzer, wherein the vector network analyzer utilizes microwave sensing for determining the plurality of reflection parameters; extract a plurality of physics-based electromagnetic (EM) features using a predetermined resonant frequency and a corresponding value of the plurality of reflection parameter at the predetermined resonant frequency to obtain a feature vector, wherein the feature vector comprises a training feature vector and a testing feature vector; classify the sample under test into one of (i) a first oedema severity category and (ii) a second oedema severity category, using a trained classifier model for the testing feature vector, wherein the trained classifier model is trained using the training feature vector; and quantify the oedema based on a water percentage in the sample under test that is classified into the first oedema severity category and the second oedema severity category, using a trained regression model for a specific set of testing feature vector from the testing feature vector.
8. The system as claimed in claim 7, wherein the sample under test is a biological tissue.
9. The system as claimed in claim 7, wherein the specific operating frequency of the tuned antenna is 4 Giga Hertz (GHz) and the predefined distance of the tuned antenna from the numerical human phantom is 10 millimeter (mm).
10. The system as claimed in claim 7, wherein the plurality of physics-based electromagnetic (EM) features specify water accumulation in the sample under test that represents a detected oedema.
11. The system as claimed in claim 7, wherein the first oedema severity category represents a low oedema severity category, the second oedema severity category represents a medium oedema severity category.
12. The system as claimed in claim 7, wherein the step of classifying comprises classifying the sample under test into a third oedema severity category that represents a low oedema severity category.
13. One or more non-transitory machine-readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors cause: acquiring a plurality of scan measurements at a plurality of locations of a sample under test using a tuned antenna, wherein the tuned antenna is a microstrip patch antenna operating at a specific operating frequency with one or more optimized dimensions and positioned at a predefined distance from a numerical biological phantom; determining a plurality of reflection parameters from the plurality of scan measurements at each location from the plurality of locations of the sample under test using a vector network analyzer, wherein the vector network analyzer utilizes microwave sensing for determining the plurality of reflection parameters; extracting a plurality of physics-based electromagnetic (EM) features using a predetermined resonant frequency and a corresponding value of the plurality of reflection parameters at the predetermined resonant frequency to obtain a feature vector, wherein the feature vector comprises a training feature vector and a testing feature vector; classifying the sample under test into one of (i) a first oedema severity category, and (ii) a second oedema severity category, using a trained classifier model for the testing feature vector, wherein the trained classifier model is trained using the training feature vector; and quantifying the oedema based on a water percentage in the sample under test that is classified into the first oedema severity category and the second oedema severity category, using a trained regression model for a specific set of testing feature vector from the testing feature vector.
14. The one or more non-transitory machine-readable information storage mediums as claimed in claim 13, wherein the sample under test is a biological tissue.
15. The one or more non-transitory machine-readable information storage mediums as claimed in claim 13, wherein the specific operating frequency of the tuned antenna is 4 Giga Hertz (GHz) and the predefined distance of the tuned antenna from the numerical human phantom is 10 millimeter (mm).
16. The one or more non-transitory machine-readable information storage mediums as claimed in claim 13, wherein the plurality of physics-based electromagnetic (EM) features specify water accumulation in the sample under test that represents a detected oedema.
17. The one or more non-transitory machine-readable information storage mediums as claimed in claim 13, wherein the first oedema severity category represents a low oedema severity category, and the second oedema severity category represents a medium oedema severity category.
18. The one or more non-transitory machine-readable information storage mediums as claimed in claim 13, wherein the step of classifying comprises classifying the sample under test into a third oedema severity category that represents a low oedema severity category.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS AND PRIORITY The present application claims priority to Indian application no. 202421085558, filed on November 07, 2024. TECHNICAL FIELD The disclosure herein generally relates to microwave sensing, and, more particularly, to a system and method for unobtrusive oedema severity classification and quantification using microwave sensing. BACKGROUND Oedema is accumulation of water in interstitial spaces beneath skin. It is not a disease in itself, rather occurs due to various underlying conditions including heart, kidney, liver diseases, certain medications and many more. Oedema monitoring and intervention thus becomes essential to effectively treat the underlying condition(s). However, typically oedema is only detected when water accumulation increases to such a great extent that swelling (i.e., physical deformation) occurs in peripheral areas (e.g., lower hands and/or legs). Hence, quantitative monitoring of oedema severity even before its physical manifestation becomes crucial. Most frequently used method of clinical assessment of oedema is either visual inspection or pitting oedema scale. However, both these methods are prone to subjective assessment error. Other than these methods, imaging techniques like Lymphoscintigraphy, ultrasonography, and magnetic resonance imaging (MRI) can evaluate oedema, but they are not portable, use harmful radiation, and are costly with long preparation time. It is observed microwave based method has lot of potential for oedema severity detection. This makes it possible to monitor variations in tissue hydration levels, such as those associated with oedema. Advantages of using microwave technology include its non-invasive nature, potential for continuous monitoring, and ability to provide information about tissue properties at different depths. However, with microwave-based assessment of oedema, it becomes challenging to gather information about the tissue properties since air-skin interface reflects most of incident microwave energy. Current microwave-based oedema or tissue hydration detection methods mostly rely on open-ended coaxial probe measurements, which are contact-based in nature and are incapable of assessing the extent of oedema typically beyond 2-3 mm depth. Other existing methods that have explored oedema assessment utilize either imaging-based approaches or contact-based sensors, which are unsuitable for efficient long-term monitoring. Microwave-based biomedical sensing suffers from a few critical challenges like antenna impedance mismatch due to proximity to biological medium, sensing resolution and penetration depth. For greater penetration and improved spatial resolution, the antenna should be placed near the biological media which in-turn affects a shift in resonance frequency of the antenna. Considering these impediments, it appears that conventional modalities for oedema detection either suffer from subjective assessment errors or involve setups that require harmful radiation, are expensive, tedious, and bulky. SUMMARY Embodiments of the present disclosure present technological improvements as solutions to one or more of the above-mentioned technical problems recognized by the inventors in conventional systems. For example, in one aspect, a processor implemented method is provided. The processor implemented method comprising acquiring, via one or more hardware processors, a plurality of scan measurements at a plurality of locations of a sample under test using a tuned antenna, wherein the tuned antenna is a microstrip patch antenna operating at a specific operating frequency with one or more optimized dimensions and positioned at a predefined distance from a numerical biological phantom; determining, via the one or more hardware processors, a plurality of reflection parameters from the plurality of scan measurements at each location from the plurality of locations of the sample under test using a vector network analyzer, wherein the vector network analyzer utilizes microwave sensing for determining the plurality of reflection parameters; extracting, via the one or more hardware processors, a plurality of physics-based electromagnetic (EM) features using a predetermined resonant frequency and a corresponding value of the plurality of reflection parameters at the predetermined resonant frequency to obtain a feature vector, wherein the feature vector comprises a training feature vector and a testing feature vector; classifying, via the one or more hardware processors, the sample under test into one of (i) a first oedema severity category, and (ii) a second oedema severity category, using a trained classifier model for the testing feature vector, wherein the trained classifier model is trained using the training feature vector; and quantifying, via the one or more hardware processors, the oedema based on a water percentage in the sample under test that is classified into the first oedema severity category and the second o