Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
CBD-NET-BASED MEDICAL ENDOSCOPIC IMAGE DENOISING METHOD
Document Type and Number:
WIPO Patent Application WO/2022/022494
Kind Code:
A1
Abstract:
Disclosed is a CBD-Net-based medical endoscopic image denoising method. The method comprises: constructing and improving a CBD-Net to obtain a medical endoscopic image denoising model; obtaining a noise source according to an imaging principle of an endoscope; acquiring a simulated training set and a real training set, and acquiring a supplementary training set according to a noise generation source; training the medical endoscopic image denoising model by using the simulated training set, the real training set and the supplementary training set; and denoising a medical endoscopic image by using the trained medical endoscopic image denoising model. In the present invention, an endoscopic image is processed without changing factors such as an original tone and the brightness thereof; and by means of the present invention, medical image details are better retained.

Inventors:
WANG YANGANG (CN)
ZHANG CHENGTIAN (CN)
CHEN YANG (CN)
Application Number:
PCT/CN2021/108608
Publication Date:
February 03, 2022
Filing Date:
July 27, 2021
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
NANJING TUGE HEALTHCARE CO LTD (CN)
International Classes:
G06T5/00; G06T7/00
Foreign References:
CN111862060A2020-10-30
CN110766632A2020-02-07
CN111192215A2020-05-22
Other References:
GUO SHI; YAN ZIFEI; ZHANG KAI; ZUO WANGMENG; ZHANG LEI: "Toward Convolutional Blind Denoising of Real Photographs", 2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), IEEE, 15 June 2019 (2019-06-15), pages 1712 - 1722, XP033687535, DOI: 10.1109/CVPR.2019.00181
LIU SHIQI: "Study of Navigation System for Percutaneous Coronary Intervention", CHINA MASTER’S THESES FULL-TEXT DATABASE), no. 8, 30 August 2019 (2019-08-30), XP055890916
See also references of EP 4187484A4
Attorney, Agent or Firm:
BEIJING FINELAND IP FIRM (CN)
Download PDF: