Copy-Move Forgery Detection Optimization Using Harmony Search with Clonal Selection Algorithm

In today's world, where technology is advancing rapidly, the extensive use of digital images on social media brings up concerns about their authenticity. One common form of deception is image forgery, especially copy-move forgery, where parts of an image are duplicated and inserted elsewhere within the same image. Traditional detection methods, like randomized search algorithms, often struggle with adaptability. This study presents a new approach that uses the Harmony Search with Clonal Selection Algorithm (HSCSA) to improve the detection of copy-move forgeries by fine -tuning the hyperparameters of Convolutional Neural Network (CNN) models. Our method involves preprocessing the CoMoFoD dataset, optimizing essential hyperparameters using HSCSA, and comparing the performance of the optimized model with standard CNN models. The results showed a notable improvement, with accuracy rising from 67.72% to 78.20%.