The test step of the Digital Anastylosis of Frescoes challenge

As one of the oldest forms of human painting, murals have rich and colorful content, making it an important research material for political history, cultural history, and art history. However, the murals are immovable cultural relics, and they are gradually aging due to weathering, wind erosion, fading, etc., they may even be destroyed by natural disasters such as earthquakes. The in-situ technique of image fragmentation is conducive to future reconstruction. And the mural information can be saved permanently. This international challenge is to contribute to the preservation of historical heritage. The goal is to identify a suitable virtual reconstruction tool for the subsequent stages of reconstruction. Anastylosis is a technique of restoration of the destroyed site. The goal is to develop tools enabling digital reconstruction of heavily damaged frescoes, supporting their restoration through the solution of the ‘puzzles’ formed by the remaining fragments, often mixed with spurious elements. The reconstruction will be made by recombining the various fragments collected and cataloged after the collapse. It will be necessary to collect all the possible data on the original artwork as the information initially available will concern only aspects such as color composition and shape. There are 3 Critical aspects： 1.Lack of continuity (some pieces have gone irretrievably lost) 2.The boundaries of the collected pieces do not match 3.There may be interference (in the same group there may be pieces of different frescoes) This paper mainly compares the output of the virtual reconstruction tool with the original image, and judges which virtual reconstruction tool is the best. We want to compare the color, area, coordinates and other factors of the result with the original image. Finally, we use the algorithm to select the optimal virtual reconstruction tool.