Title:
IMAGE NOISE LEARNING SERVER AND IMAGE NOISE REDUCTION DEVICE USING MACHINE LEARNING
Document Type and Number:
WIPO Patent Application WO/2024/010133
Kind Code:
A1
Abstract:
An image noise learning server comprises: an image input unit which receives an input of a training image; an image extraction unit which extracts a first image including a still object from the input image and extracts a second image including a moving object from the input image; a noise filtering unit which obtains a third image by applying a noise filter having a first intensity to the second image; a labeling unit which calculates the size of a side effect by using a difference between the still object included in the first image and a moving object included in the third image; a machine learning unit which obtains an AI parameter by performing machine learning for the second image by inputting the calculated size of the side effect and image properties of the input image by using a label; and a communication unit which transmits the obtained AI parameter.
Inventors:
KIM DAE BONG (KR)
PARK DONG JIN (KR)
PARK DONG JIN (KR)
Application Number:
PCT/KR2022/012508
Publication Date:
January 11, 2024
Filing Date:
August 22, 2022
Export Citation:
Assignee:
HANWHA VISION CO LTD (KR)
International Classes:
G06T5/00; G06N3/08; G06N20/00; G06T5/20; G06T7/11; G06V20/70; G06V40/10
Foreign References:
| KR101987079B1 | 2019-06-10 | |||
| KR20170058277A | 2017-05-26 | |||
| KR20100020068A | 2010-02-22 | |||
| US20210019507A1 | 2021-01-21 | |||
| US20190333197A1 | 2019-10-31 |
Attorney, Agent or Firm:
KASAN IP & LAW FIRM et al. (KR)
Download PDF: