Abstract: With the ease with which images may be altered using advanced software, digital image fraud has become a widespread problem in the modern digital era. To maintain the integrity and authenticity of digital photographs, the detection of such forgeries is essential in a number of sectors, including journalism, forensics, and legal investigations. The intricacy and subtlety of contemporary forgeries can provide a challenge to conventional image forgery detection techniques. In this study, we provide a deep learning-based method for efficiently identifying digital image forgeries. Using convolutional neural networks (CNNs), our approach automatically learns and extracts information from photos, making it possible to identify many kinds of forgeries, including copy-move,  slicing, and image retouching to improve its accuracy and robusteness. suggested system is trained on a variety of forged and actual images. Our deep learning-based methodology works better than conventional methods, as evidenced by experimental results that show lower false positive rates and higher detection rates. By offering a potent tool for the automatic detection of image forgeries, this research advances the field of digital image forensics and ensures the authenticity and dependability of digital visual content.

Keywords: CNN, Forgery, ELA

| DOI: 10.17148/IARJSET.2024.11746