Image fragment reconstruction and restoration method based on genetic algorithm

Abstract

How to efficiently and successfully stitch together fragments to reformat them into original images is an important and challenging problem in many disciplines, especially forensic science, archaeology, and investigative science. It is challenging for humans to perform this task as it requires much time, skill, and effort, and humans are prone to make mistakes. Therefore, automating the process through appropriate technology can significantly increase the problem-solving rate. The task can be accomplished faster and more efficiently, reducing the amount of work required to stitch the fragments together while increasing the efficiency of the entire process and significantly reducing human involvement. This study proposes an image restoration method based on reconstructing the edges of torn fragments. Torn fragments are often lost due to how they were initially torn, so an innovative model is proposed to replace the missing parts, i.e., using the Fourier inverse in combination with a genetic algorithm for edge replacement. For validation, 17 fragments of a particular image were input. The success rate of reconstruction was 86.6% for 15 of these images, 100% for 5 images, and 87.5% for the combined success rate, which proves the validity of the hypothesis of using the Fourier inverse to reconstruct the edges.