Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
METHOD FOR TRAINING IMAGE QUALITY IMPROVEMENT MODEL, AND METHOD FOR IMPROVING IMAGE QUALITY OF VIDEO CONFERENCE SYSTEM
Document Type and Number:
WIPO Patent Application WO/2023/202447
Kind Code:
A1
Abstract:
The embodiments of the present application relate to the technical field of computer vision. Disclosed are a method for training an image quality improvement model, and a method for improving the image quality of a video conference system. The image quality improvement model is used for improving the image quality of a video image of a video conference system. The training method comprises: acquiring at least one noise model, wherein the noise model is obtained by means of performing training on the basis of a first noisy image, and is used for simulating noise characteristics of a video conference system; acquiring a noise-free image set, and inputting a noise-free image in the noise-free image set into the noise model, so as to obtain a second noisy image, wherein the noise-free image and the second noisy image form a training data pair; and training an initial image quality improvement model by using the training data pair, and after training is completed, performing convolution folding on the trained image quality improvement model, so as to obtain a final image quality improvement model.

Inventors:
XU QIAN (CN)
YU WEINA (CN)
Application Number:
PCT/CN2023/087910
Publication Date:
October 26, 2023
Filing Date:
April 12, 2023
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
ZTE CORP (CN)
International Classes:
G06T5/00
Domestic Patent References:
WO2022012888A12022-01-20
Foreign References:
CN107993200A2018-05-04
CN108229525A2018-06-29
CN113628146A2021-11-09
CN113822289A2021-12-21
Attorney, Agent or Firm:
COHORIZON INTELLECTUAL PROPERTY INC. (CN)
Download PDF: