Abstract:Mural is an important component of culture and art of Dunhuang in China. Unfortunately, these murals had been ruined or are being ruined by some diseases such as cracking, hollowing, falling off, mildew, dirt, and so on. Existing image restoration algorithms have problems such as incomplete repair and disharmonious texture during large-area repair, so the effect of mural image disease area repair is poor. Due to lack of a standard mural datasets, Dunhuang mural datasets are created in the paper. Meanwhile, our network architecture SeparaFill is proposed which connects two generators based on U-Net. Based on the characteristics of the painting, the contour line pixel area of the mural image is innovatively separated from the content pixel area. Firstly, the contour restoration generator network with skip connect and hierarchical residual blocks is employed to repair contour lines. Then, the color mural image is repaired by the content completion network with guide of the repaired contour. Full resolution branches and generator branches of the U type are exploited in content completion generators. Convolution layers of different kernel sizes are fused to improve the reusability of the underlying features. Finally, global and local discriminant networks are applied to determine whether the repaired mural image is authentic in terms of both the modified and unmodified areas. The proposed SeparaFill shows good performance in restoring the line structure of the damaged mural images and retaining the contour information of the mural images. Compared with existing restoration algorithms in mural real damage repair experiment, our algorithm increases the peak signal-to-noise ratio (PSNR) by an average of 1.1–4.3 dB and the structural similarity (SSIM) values were slightly improved. Experimental results reveal the good performance of the proposed model, which can contribute to digital restorations of ancient murals.

Multi-stage Feature Fusion Network for Edge Detection of Dunhuang Murals

Dunhuang Mural Line Drawing Based on Multi-scale Feature Fusion and Sharp Edge Learning

Dunhuang Mural Line Drawing Based on Bi-Dexined Network and Adaptive Weight Learning.

Dunhuang murals contour generation network based on convolution and self-attention fusion

Progressive Dunhuang Murals Inpainting Based on Attention Mechanism

Knowledge Based Lacunas Detection And Segmentation For Ancient Paintings

Localizing Artistic Objects in Frescos Through Contour Based Features

Progressive Dunhuang murals inpainting based on recurrent feature reasoning network

Dunhuang murals image restoration method based on generative adversarial network

Two-stage mural image restoration using an edge-constrained attention mechanism

Dunhuang mural inpainting based on reference guidance and multi‐scale fusion

Multi-stage Progressive Reasoning for Dunhuang Murals Inpainting

Inpainting Digital Dunhuang Murals with Structure-Guided Deep Network

Detection of Damaged Areas in Ancient Murals Based on Multi-Scale Dual Channel Attention U-Net

A Double Feature Fusion Network with Progressive Learning for Sharper Inpainting

Wavelet-Based Dual-Stream Network for Inpainting of Dunhuang Murals

Learning Multi-Path CNN for Mural Deterioration Detection

SeparaFill: Two generators connected mural image restoration based on generative adversarial network with skip connect

DunHuangStitch: Unsupervised Deep Image Stitching of Dunhuang Murals

Edge Detection via Fusion Difference Convolution

Restoration of Dunhuang Murals on Large-scale pretraining