Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
SYSTEM AND METHOD FOR PREDICTING DEPTH IMAGE FOR MEDICAL IMAGE
Document Type and Number:
WIPO Patent Application WO/2023/113500
Kind Code:
A1
Abstract:
Disclosed are a system for predicting a depth image for a medical image and a control method thereof. A specific embodiment of the present technology involves segmenting a collected medical image and generating a three-dimensional medical image in the form of a three-dimensional volume, converting the generated three-dimensional medical image into a virtual medical image in a real space or a realistic space and then respectively deriving a virtual medical image and a depth image from the converted virtual medical image, training a neural network on the texture and color of the derived virtual medical image and converting an image domain, training the neural network on the domain-converted virtual medical image and outputting a depth image, and storing the pair of outputted depth image and virtual medical image as a correct answer image and building a learning model therefrom, so as to predict a depth image for an inputted medical image on the basis of a neural network based on a learning model, derive the location of a lesion accurately from a high-resolution depth image, and thereby increase the rate of diagnosing a lesion.

Inventors:
SON YOUNG-DON (KR)
KIM HANG-KEUN (KR)
KANG CHANG-KI (KR)
JEONG BONG-HYUK (KR)
Application Number:
PCT/KR2022/020457
Publication Date:
June 22, 2023
Filing Date:
December 15, 2022
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
UNIV GACHON IND ACAD COOP FOUND (KR)
International Classes:
G06T7/50; G06T15/04; G06T17/20; G06T19/00; G16H30/00; G16H50/50
Foreign References:
KR101608848B12016-04-04
KR102245220B12021-04-27
KR101959438B12019-03-18
KR20150094080A2015-08-19
JP6931425B22021-09-01
Attorney, Agent or Firm:
LEE, Un Cheol (KR)
Download PDF: