Method for providing necessary information for diagnosis of Alzheimer's disease from photographed image of patient using deep learning and apparatus for executing the method

Patent metadata

Field	Value
Publication number	US-12462607-B2
Application number	US-202217972057-A
Country	US
Kind code	B2
Filing date	Oct 24, 2022
Priority date	Oct 25, 2021
Publication date	Nov 4, 2025
Grant date	Nov 4, 2025

Abstract

Disclosed are a method for providing the necessary information for a diagnosis of Alzheimer's disease and an apparatus for executing the method. The apparatus for executing the method for providing the necessary information for a diagnosis of Alzheimer's disease includes one or more processors, a memory, and one or more programs, in which the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs includes an instruction for acquiring a photographed image of a patient's eyeball, an instruction for preprocessing the photographed image, generating a blood vessel image from the pre-processed photographed image using machine learning-based technology, and providing the necessary information for a diagnosis of Alzheimer's disease based on the generated blood vessel image, and an instruction for generating diagnostic prediction information based on the necessary information for a diagnosis of Alzheimer's disease.

First claim

What is claimed is:

1 . A computing device comprising:
one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs include: an instruction for acquiring a photographed image obtained by photographing a patient's eyeball, an instruction for preprocessing the photographed image, generating a blood vessel image from the pre-processed photographed image using machine learning-based technology, and providing necessary information for a diagnosis of Alzheimer's disease based on the generated blood vessel image, and an instruction for generating diagnostic prediction information based on the necessary information for a diagnosis of Alzheimer's disease, wherein the instruction for providing the necessary information for the diagnosis of Alzheimer's disease includes: an instruction for generating a data set by extracting a retinal region from the photographed acquired image, and an instruction for classifying a patient's risk rating for Alzheimer's disease from the data set by receiving the data set using a machine learning module, wherein the instruction for classifying the patient's risk rating for Alzheimer's disease includes: an instruction for generating a feature map from a retinal image included in the data set by receiving the data set using an encoder module, an instruction for generating the blood vessel image based on the feature map by receiving the feature map using a decoder module, an instruction for generating a final blood vessel image based on the retinal image, the feature map, and the blood vessel image by receiving the retinal image, the feature map, and the blood vessel image using a generation module, and an instruction for performing category classification to provide the necessary information for the diagnosis of Alzheimer's disease based on the final blood vessel image by receiving the final blood vessel image using a classification module, wherein the instruction for classifying the patient's risk rating for Alzheimer's disease further includes: an instruction for generating a feature map from a retinal image included in the data set by receiving the data set using an encoder module, an instruction for generating the blood vessel image based on the feature map by receiving the feature map using a decoder module, an instruction for generating a final blood vessel image based on the retinal image, the feature map, and the blood vessel image by receiving the retinal image, the feature map, and the blood vessel image using a generation module, and an instruction for performing category classification to provide the necessary information for the diagnosis of Alzheimer's disease based on the final blood vessel image by receiving the final blood vessel image using a classification module, wherein the generation module is trained to: divide the feature map and the blood vessel image into the same size, respectively, by receiving the retinal image, the feature map, and the blood vessel image, generate a similarity distribution by comparing similarities between any one of a plurality of the divided blood vessel images and a plurality of the divided feature maps, respectively, reflect the similarity distribution in the retinal image, and extract a blood vessel from a region corresponding to any one of the plurality of divided blood vessel images from the retinal image using the retinal image in which the similarity distribution is reflected and any one of the plurality of divided blood vessel images.

2 . The device of claim 1 , wherein
the instruction for generating the data set further includes an instruction for generating a retinal image by extracting a retinal region from the acquired photographed image, an instruction for adjusting the retinal image to a preset size, and an instruction for normalizing the adjusted retinal image.

3 . The device of claim 1 , wherein
the encoder module is trained to generate a feature map through a convolution operation while moving a first filter to the retinal image by receiving the retinal image.

4 . The device of claim 1 , wherein
the decoder module is trained to generate a blood vessel image from the feature map through a deconvolution operation while moving a second filter to the feature map by receiving the generated feature map.

5 . The device of claim 1 , wherein
the instruction for generating the final blood vessel image further includes an instruction for repeating so that blood vessels are extracted from all regions of the retinal image by the generation module, and an instruction for generating a final blood vessel image by using the blood vessels extracted from all regions of the retinal image.

6 . The device of claim 1 , wherein
the classification module is trained to classify a risk rating for Alzheimer's disease based on a preset reference image and the generated final blood vessel image by receiving the final blood vessel image.

7 . A method for providing necessary information for a diagnosis of Alzheimer's disease performed by a computing device that includes one or more processors and a memory storing one or more programs executed by the one or more processors, the method comprising:
acquiring a photographed image obtained by photographing a patient's eyeball; preprocessing the photographed image, generating a blood vessel image from the pre-processed photographed image using machine learning-based technology, and providing the necessary information for a diagnosis of Alzheimer's disease based on the generated blood vessel image; and generating diagnostic prediction information based on the necessary information for the diagnosis of Alzheimer's disease, wherein the providing of the necessary information for the diagnosis of Alzheimer's disease includes: generating a data set by extracting a retinal region from the acquired photographed image, and classifying a patient's risk rating for Alzheimer's disease from the data set by receiving the data set using a machine learning module, wherein the classifying of the patient's risk rating for Alzheimer's disease includes: generating a feature map from a retinal image included in the data set by receiving the data set using an encoder module, generating the blood vessel image based on the feature map by receiving the feature map using a decoder module, generating a final blood vessel image based on the retinal image, the feature map, and the blood vessel image by receiving the retinal image, the feature map, and the blood vessel image using a generation module, and performing category classification to provide the necessary information for the diagnosis of Alzheimer's disease based on the final blood vessel image by receiving the final blood vessel image using a classification module, and wherein the generation module is trained to: divide the feature map and the blood vessel image into the same size, respectively, by receiving the retinal image, the feature map, and the blood vessel image, generate a similarity distribution by comparing similarities between any one of a plurality of the divided blood vessel images and a plurality of the divided feature maps, respectively, reflect the similarity distribution in the retinal image, and extract a blood vessel from a region corresponding to any one of the plurality of divided blood vessel images from the retinal image using the retinal image in which the similarity distribution is reflected and any one of the plurality of divided blood vessel images.

8 . The method of claim 7 , wherein
the generating the data set further includes generating a retinal image by extracting a retinal reg

Assignees

• Univ Ajou Ind Academic Coop Found

Inventors

• Kim Do Young
• Sunwoo Myung Hoon
• Lim Young Jun
• Park Joon-Hyeon

Classifications

• G06T2207/30101

  Blood vessel; Artery; Vein; Vascular · CPC title

• G06T2207/30041

  Eye; Retina; Ophthalmic · CPC title

• G06T7/0012

  Biomedical image inspection · CPC title

• G06V2201/07

  Target detection · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title