Bring old mural images to life
Wenyuan Xi, Eryu Wang, Ziyin Shen · 2023
This research focuses on the detection and digital restoration of cracks in ancient frescoes using a CNN model. The proposed method effectively preserves the detailed information of the frescoes and minimizes damage to the original artwork during the restoration process. The CNN model is employed to digitally repair the mural and optimize the repair performance by adjusting the parameters to achieve a natural-looking restoration of any cracks. Experimental results demonstrate the notable advantages of the proposed model in the digital restoration of murals and provide an effective method for the conservation and preservation of ancient frescoes. This study also holds significant implications for the technological advancements in the field of mural restoration and heritage conservation, contributing to progress and innovation in related areas.