Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
SUPER-RESOLUTION MEDICAL IMAGE GENERATION APPARATUS AND METHOD
Document Type and Number:
WIPO Patent Application WO/2022/239897
Kind Code:
A1
Abstract:
A super-resolution medical image generation apparatus (100) according to the present invention comprises: a data set management unit (200) for managing a data set (210) including multiple images including a high-resolution original image (211), a low-resolution conversion image (212), a primary super-resolution conversion image (213), a secondary super-resolution conversion image (214), a low-resolution original image (215), and a super-resolution reconstruction image (216); a training model (300) including a generator (310) and a discriminator (320); and a training unit (400) for training the generator (310) through prior training and training the generator (310) and the discriminator (320) through main training which proceeds after the prior training, wherein the data set management unit (200) comprises a preprocessing unit (220) for converting the high-resolution original image (211) into the low-resolution conversion image (212).

Inventors:
LEE WON SEOP (KR)
AHN JAE HYUN (KR)
PARK GANG IN (KR)
SONG MYUNG SEO (KR)
Application Number:
PCT/KR2021/006386
Publication Date:
November 17, 2022
Filing Date:
May 21, 2021
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
CNAI CO LTD (KR)
International Classes:
G06T3/40
Domestic Patent References:
WO2019209820A12019-10-31
Foreign References:
KR20190040586A2019-04-19
KR20200101493A2020-08-28
KR20200084434A2020-07-13
Other References:
SEO JEONG-IN: "Very Low Resolution Object Recognition using Deep Collaborative Learning of Image Enhancement and Classification", PHD THESIS, 1 August 2019 (2019-08-01), KR, pages 1 - 93, XP009541077
Attorney, Agent or Firm:
SHIN, Moo Yun (KR)
Download PDF: