Research on Image Super-Resolution Reconstruction Algorithm Based on Latent Auxiliary Feature
As one of the basic research tasks in image quality enhancement field,image super-resolution has high research and appli-cation value.Generative adversarial network can effectively improve the texture detail information of super-resolution images,which has been widely used in this field.However,it is difficult to generate high-quality images relying on the features just learned from the input low-resolution images.To solve this problem,this paper proposes an image super-resolution reconstruction algorithm based on latent auxiliary feature,which introduces a trainable latent feature to expand the feature space of the generator,and provides auxiliary feature information to improve visual performance.Moreover,we also utilize the image feature to constrain the latent feature space,which can avoid low fidelity of the super-resolution image content.The proposed method was compared with seven methods on seven public datasets.Experimental results show that our proposed method has richer detail information and better visual effect.
NETL
NSTL
万方数据
