Innovations in Image Forensics: Designing a Smart Learning Based Methodology for Forgery Detection over Digital Images
G. Ramkumar, Khobragade Pithamber, Ch. Srilakshmi, S. Rathika, B. Anni Princy, Vijay Anand Kandaswamy
Abstract
Detecting forgery in digital images is crucial for several reasons. Firstly, with the widespread use and accessibility of image editing tools, individuals can easily manipulate digital images, raising concerns about the authenticity and integrity of visual information. This study introduces a novel approach, termed SFVGG, for detecting forgery in digital images by leveraging a smart learning methodology. The SFVGG method integrates the robust features of SIFT, the clustering capabilities of Fuzzy C-Mean clustering, and the deep learning process of VGG19. Through a synergistic combination of these techniques, the proposed forgery detection system achieves a comprehensive and effective solution. The methodology was implemented using Python, showcasing its practical applicability. Remarkably, the SFVGG framework demonstrates a high accuracy of 94.5%, underscoring its effectiveness in identifying forged content within digital images. This research represents a significant advancement in digital image forgery detection, offering both a sophisticated methodology and practical implementation.