Title:
OSTEOPOROSIS PREDICTION SYSTEM USING DEEP LEARNING MODULE
Document Type and Number:
WIPO Patent Application WO/2024/080428
Kind Code:
A1
Abstract:
The present invention relates to an osteoporosis prediction system, comprising: a preprocessing module; and first to n-th deep learning modules (where n is a natural number greater than or equal to 2), wherein the preprocessing module filters a CT image of the skeleton so as to have hounsfield units (HU) of different bands to generate first to n-th filtered CT images, and the first to n-th deep learning modules learn the first to n-th filtered CT images, respectively, by using a dissimilarity loss function that makes the similarity of feature information of the first to n-th deep learning modules small.
Inventors:
HWANG DONG HWAN (KR)
KIM YOON (KR)
CHOI HYUN-SOO (KR)
BAK SO HYEON (KR)
KIM YOON (KR)
CHOI HYUN-SOO (KR)
BAK SO HYEON (KR)
Application Number:
PCT/KR2022/016124
Publication Date:
April 18, 2024
Filing Date:
October 21, 2022
Export Citation:
Assignee:
ZIOVISION CO LTD (KR)
International Classes:
A61B6/00; A61B6/03; G16H30/40; G16H50/20; G16H50/50
Foreign References:
| KR20220040437A | 2022-03-30 | |||
| KR20200062538A | 2020-06-04 | |||
| US20200160515A1 | 2020-05-21 | |||
| US20190239843A1 | 2019-08-08 |
Other References:
CHOI, Se-Heon et al. Efficient osteoporosis prediction using a pair of ensemble models. Journal of The Korea Society of Computer and Information. 31 December 2021 (publication date), vol. 26, no. 12, pp. 45-52.
Attorney, Agent or Firm:
LIM & JEONG PATENT LAW FIRM (KR)
Download PDF: