Advances in Copy Move Forgery Detection in Digital Images: A Comparative Examination of Conventional Approaches and Deep Learning Based Models

In recent years, the ease of digital image alteration has significantly increased, primarily driven by the advancements in computing technology and the availability of user-friendly professional image editing software. Consequently, the process of authenticating digital images has become increasingly complex, especially when attempting to distinguish between altered and genuine images with the naked eye. This paper aims to explore various techniques used to detect Copy-Move Forgery in digital images. It provides a comparative analysis of conventional approaches, such as Binary Discriminative Features with Mean Colour Similarity, Color Histogram, DCT and SIFT against the deep learning models AlexNet and GoogleNet. Benchmark datasets such as MICC-F2000 and CoMoFoD are used to evaluate the performance of the comparing models. While conventional techniques exhibit respectable precision, recall, and F1 scores, contemporary deep learning methods outperform them marginally.