Reference-Based Super-Resolution Reconstruction of Mural Images
Yushan Chen, Zhigang Xu · 2023
To solve the problem of texture defect and content ambiguity of degraded mural images, in this paper, we propose a reference-based image super-resolution(RefSR) method, The repaired high-resolution mural images are used as reference images, and the multi-scale information of the images is used to establish the relationship between the reference images and the low-quality mural images. Aiming at the problems of color deviation and texture authenticity that may occur in the reconstruction process, an adaptive texture transfer module and a self-tuning pixel attention module are constructed; In addition, in order to reduce the complexity and applicability of the model, residual distillation module is used to reduce the training parameters of the model.