Effective structure keeping for generative adversarial networks for single image super resolution
US11048974B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Aug 5, 2019 |
| Grant date | Jun 29, 2021 |
| Priority date | — |
| Expiry date | Dec 11, 2039 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06V10/82
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
A method of training a generator G of a Generative Adversarial Network (GAN) includes generating a real contextual data set {x1, . . . , xN} for a high resolution image Y; generating a generated contextual data set {g1, . . . , gN} for a generated high resolution image G(Z); calculating a perceptual loss Lpcept value using the real contextual data set {x1, . . . , xN} and the generated contextual data set {g1, . . . , gN}; and training the generator G using the perceptual loss Lpcept value. The generated high resolution image G(Z) is generated by the generator G of the GAN in response to receiving an input Z, where the input Z is a random sample that corresponds to the high resolution image Y.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.