Title:
IMAGE RECONSTRUCTION MODEL GENERATION METHOD, IMAGE RECONSTRUCTION METHOD, AND RELATED DEVICE
Document Type and Number:
WIPO Patent Application WO/2022/127496
Kind Code:
A1
Abstract:
Embodiments of the present application relate to the field of artificial intelligence, and provide an image reconstruction model generation method, an image reconstruction method, and a related device. The method comprises: obtaining training data; inputting the training data into a preset model for feature extraction to obtain feature data; performing nonlinear mapping on the feature data to obtain mapping data; performing up-sampling on the mapping data to obtain reconstruction data; calculating loss on the basis of the reconstruction data and corresponding preset label data, wherein the loss comprises gradient loss between the reconstruction data and the corresponding preset label data; and performing iterative training on the preset model on the basis of the loss to obtain an image reconstruction model. According to the method provided by the embodiments of the present application, the sensitivity of the model to high-frequency information such as edges can be improved, so that the quality of a reconstructed image can be improved.
Inventors:
WU JIE (CN)
LI WENGUO (CN)
LI WENGUO (CN)
Application Number:
PCT/CN2021/131268
Publication Date:
June 23, 2022
Filing Date:
November 17, 2021
Export Citation:
Assignee:
SPREADTRUM COMMUNICATIONS TIANJIN CO LTD (CN)
International Classes:
G06T3/40
Foreign References:
| CN112801866A | 2021-05-14 | |||
| CN111932460A | 2020-11-13 | |||
| CN109978762A | 2019-07-05 | |||
| CN110136063A | 2019-08-16 | |||
| US20100086227A1 | 2010-04-08 |
Attorney, Agent or Firm:
UNI-INTEL PATENT AND TRADEMARK LAW FIRM (CN)
Download PDF: