07-04-2023 дата публикации
Номер: CN115936985A
Принадлежит:
The invention discloses an image super-resolution reconstruction method based on a high-order degradation cyclic generative adversarial network, and the method comprises the steps: generating a low-resolution image for a high-resolution image through employing a high-order image degradation model, forming an image pair, constructing a cyclic generative network model, and carrying out the reconstruction of a high-order degradation cyclic generative adversarial network. The method comprises the following steps of: generating a low-resolution image to a high-resolution image by utilizing a generator, judging, constructing a network loss function, ensuring the stability of an established mapping relation through a classic GAN network loss function and a cyclic invariant Cycle loss function, finally training a network model, configuring various hyper-parameters, and establishing a high-resolution image. And the network can be converged to obtain a better generation effect. Compared with a traditional ...
Подробнее