Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12567503-B2
Publication Date
2026-03-03
Expiration Date
Abstract
A system of reconstructing a dielectric image is provided. The system includes a data collection array to collect microwave scattering data. The system includes a machine learning device configured to receive the microwave scattering data, analyze the microwave scattering data, output a generated image based on the analyzed microwave scattering data, and identify at least one of a presence of disease, absence of disease, or one or more disease features from the generated image.
Core Innovation
The invention provides a portable system of reconstructing a dielectric image using microwave tomography/dielectrography concepts. A data collection array collects microwave scattering data, which include S-parameter data, and a machine learning device receives the microwave scattering data, analyzes the data, and outputs a generated image based on the analyzed microwave scattering data.
From the generated image, the machine learning device identifies at least one of a presence of disease, absence of disease, or one or more disease features. The approach uses incidence S-parameter data and/or scattering S-parameter data and is framed as transforming a reconstruction problem into an overdetermined, population-trained learning task using broadband frequencies.
In some implementations, the machine learning device can incorporate patient information and/or imaging modality data as ground truth, including MR, CT, or PET data. The document describes proof-of-concept in silico classification and discrimination for hemorrhagic stroke detection, including localization and size quantification, and it emphasizes non-ionizing/low energy radiofrequency operation and modular helmet integration with rapid acquisition.
Claims Coverage
The independent claims include three independent items: a system, a method, and a non-transitory computer-accessible medium. Across these independent claims, the inventive features center on collecting microwave scattering data, using a machine learning device to analyze the data, generating a dielectric image from the analyzed data, and identifying at least one of disease presence/absence or disease features, with the method and medium also covering physiological attributes.
Microwave scattering data collection for dielectric image reconstruction
The system comprises a data collection array to collect microwave scattering data.
Machine learning analysis and generated dielectric image output
A machine learning device is configured to receive the microwave scattering data, analyze the microwave scattering data, and output a generated image based on the analyzed microwave scattering data.
Disease identification from the generated dielectric image
The machine learning device is configured to identify at least one of a presence of disease, absence of disease, or one or more disease features from the generated image.
Dielectric image reconstruction procedure using a machine learning device
A method obtains microwave scattering data from a data collection array, receives the microwave scattering data by a machine learning device, analyzes the microwave scattering data by the machine learning device, and outputs a generated image based on the analyzed microwave scattering data.
Identifying presence/absence of disease, physiological attributes, or disease features from the generated image
The method identifies, by the machine learning device, at least one of a presence of disease, absence of disease, one or more physiological attributes, or one or more disease features from the generated image.
Non-transitory computer-executable instructions for machine learning dielectric image reconstruction and disease identification
A non-transitory computer-accessible medium stores computer-executable instructions that, when executed, obtain microwave scattering data from a data collection array, receive and analyze the microwave scattering data by a machine learning device, output a generated image based on the analyzed microwave scattering data, and identify by the machine learning device at least one of a presence of disease, absence of disease, one or more physiological attributes, or one or more disease features from the generated image.
Across the independent claims, the core claim coverage is the combination of microwave scattering data acquisition by a data collection array with a machine learning device that analyzes the microwave scattering data to output a generated dielectric image and then identifies disease presence/absence and/or disease features from the generated image, with the method and medium also covering physiological attributes.
Stated Advantages
Portable system with modular helmet integration and rapid acquisition.
Non-ionizing/low energy radiofrequency operation.
Improved disease detection performance is described in proof-of-concept in silico results for hemorrhagic stroke detection, including classification and discrimination outputs.
Documented Applications
In silico proof-of-concept for hemorrhagic stroke detection, including classification and discrimination such as localization and size.
Use of generated dielectric images derived from microwave scattering data to identify disease presence/absence or disease features.
Interested in licensing this patent?