Filling and restoration of cracked digitized paintings based on preserved neighbourhood pixels

• Main Authors: Farajalla Ali, Nawafil Abdulwahab, Alshaikhli, Imad Fakhri Taha
• Format: Article
• Language: English
• Published: Taylor's University 2021
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://irep.iium.edu.my/93777/8/93777_Filling%20and%20restoration%20of%20cracked%20digitized%20paintings%20based.pdf; http://irep.iium.edu.my/93777/; http://jestec.taylors.edu.my
• Institution: Universiti Islam Antarabangsa Malaysia

This paper investigates the applicability of a proposed digital image algorithm for filling cracks in digitized paintings. Filling is defined as a process of modifying/restoring damaged regions or lost areas in images. Old paintings affected by cracks suffer from decreased cultural heritage quality. Paintings that are digitized typically suffer damages from humidity, dust, smoke, and weather conditions, and the filling procedure is necessary for the restoration and preservation of this cultural heritage for the future generation. An algorithm that can automatically fill the cracks by first creating a 3x3window size to identify unwanted pixels then applying local image information from neighborhood pixel preserved from a previous procedure was proposed. The results confirmed that the proposed algorithm resulted in an excellent restored image. It is also highly accurate compared to previous algorithms, confirming its suitability for restoring old digitized paints. The algorithm was evaluated by implementing PSNR (Peak Signal-to-noise ratio) and MSE (Mean Square error), and MATLAB was used to build the code required to process and analyze the data.