Old Wall Paintings Restored Using Digital Technology

Abstract

The ancient mural pictures are virtually recreated in this study using a technique that takes a weighted average of the mean image and the original image. There are four phases to this approach. Correlation and convolution are used to detect and extract lines. Toggle filter is used to improve the extracted lines. Templates with various orientations are used for correlation and convolution. Using k-means clustering, the pixels are grouped and each pixel is replaced by the cluster's mean value. After that, the weighted average is used to restore the ancient wall paintings.