Optical coherence tomography-based system for diagnosis of high-risk lesion and diagnosis method therefor
AssigneesThe Asan Foundation, University of Ulsan Foundation for Industry Cooperation
InventorsSoo Jin Kang, Hyun Seok Min
Life of this patent
Estimating the term…
Abstract
The disclosure purposes to provide an optical coherence tomography (OCT)-based system for diagnosing a high risk lesion such as a vulnerable atheromatous plaque by using an artificial intelligence model through deep learning. A deep learning-based diagnostic method of diagnosing a high risk lesion of a coronary artery includes: acquiring an OCT image of a coronary artery lesion of a patient; extracting a first feature of a thin cap from the OCT image; setting a region of interest included in the OCT image on a basis of the first feature; and determining whether the region of interest includes a high risk lesion.
Claims11
A deep learning-based diagnostic method of diagnosing a high risk lesion of a coronary artery, the deep learning-based diagnostic method comprising:
- acquiring, by a high risk lesion diagnostic apparatus, an optical coherence tomography (OCT) image of a coronary artery lesion of a patient;
- extracting, by the high risk lesion diagnostic apparatus, a first feature of a thin cap and a second feature about a necrotic core from the OCT image using an artificial intelligence (AI) model;
- setting, by the high risk lesion diagnostic apparatus, a region of interest included in the OCT image according to whether the first feature satisfies a preset first reference;
- and
- determining, by the high risk lesion diagnostic apparatus, whether the region of interest includes a high risk lesion according to whether the second feature satisfies a preset second reference, wherein
- the AI model is trained to receive a plurality of OCT images as input data, extract a plurality of features included in the OCT image for each of the plurality of OCT images, and output a feature map indicating whether to include the high risk lesion based on the plurality of features and a result value, as output data,
- the AI model adjusts an initial learning rate by an Adam optimizer, and
- a training data set including the input data and the output data is classified according to whether the high risk lesion has been classified, and a result of the classification is evaluated and repeatedly trained.
A non-transitory computer-readable recording medium storing therein an operating program that causes a computer to execute a process comprising:
- acquiring an optical coherence tomography (OCT) image of a coronary artery lesion of a patient;
- extracting a first feature of a thin cap and a second feature about a necrotic core from the OCT image using an artificial intelligence (AI) model;
- setting a region of interest included in the OCT image according to whether the first feature satisfies a preset first reference;
- and
- determining whether the region of interest includes a high risk lesion according to whether the second feature satisfies a preset second reference, wherein
- the AI model is trained to receive a plurality of OCT images as input data, extract a plurality of features included in the OCT image for each of the plurality of OCT images, and output a feature map indicating whether to include the high risk lesion based on the plurality of features and a result value, as output data,
- the AI model adjusts an initial learning rate by an Adam optimizer, and
- a training data set including the input data and the output data is classified according to whether the high risk lesion has been classified, and a result of the classification is evaluated and repeatedly trained.
An apparatus comprising:
- processing circuitry configured to
- acquire an optical coherence tomography (OCT) image of a coronary artery lesion of a patient,
- extract a first feature of a thin cap and a second feature about a necrotic core from the OCT image using an artificial intelligence (AI) model,
- set a region of interest included in the OCT image according to whether the first feature satisfies a preset first reference, and
- determine whether the region of interest includes a high risk lesion according to whether the second feature satisfies a preset second reference, wherein
- the AI model is trained to receive a plurality of OCT images as input data, extract a plurality of features included in the OCT image for each of the plurality of OCT images, and output a feature map indicating whether to include the high risk lesion based on the plurality of features and a result value, as output data,
- the AI model adjusts an initial learning rate by an Adam optimizer, and
- a training data set including the input data and the output data is classified according to whether the high risk lesion has been classified, and a result of the classification is evaluated and repeatedly trained.
Dependencies are inferred from the claim text.