Title:
DEEP-LEARNING-BASED FUNDUS IMAGE CLASSIFICATION DEVICE AND METHOD FOR DIAGNOSIS OF OPHTHALMOLOGICAL DISEASES
Document Type and Number:
WIPO Patent Application WO/2022/085986
Kind Code:
A1
Abstract:
Disclosed are a deep-learning-based fundus image classification device and method for diagnosing ophthalmological diseases. The deep-learning-based fundus image classification device for diagnosing ophthalmological diseases, according to one aspect of the present invention, comprises: a data preprocessing unit, which preprocesses retinal fundus image data to normalize image data before model training; and a retinal image classification unit, which trains and tests the preprocessed retinal fundus image data to classify retinal images, and uses stratified 10-fold cross validation to classify retinal images.
Inventors:
NAM YUN-YOUNG (KR)
Application Number:
PCT/KR2021/013512
Publication Date:
April 28, 2022
Filing Date:
October 01, 2021
Export Citation:
Assignee:
UNIV SOONCHUNHYANG IND ACAD COOP FOUND (KR)
International Classes:
A61B3/12; A61B3/00; A61B5/00; G06N3/08; G06T7/00; G16H30/40; G16H50/20
Foreign References:
KR20200005433A | 2020-01-15 | |||
KR102150647B1 | 2020-09-02 | |||
KR20200072127A | 2020-06-22 | |||
KR101848321B1 | 2018-04-20 | |||
KR101857624B1 | 2018-05-14 |
Attorney, Agent or Firm:
WETHEPEOPLE IP & LAW FIRM (KR)
Download PDF: