FAKE IMAGE DETECTION

Abstract

we explore the potential of robust hashing techniques in effectively detecting fake 
images, even in the presence of multiple manipulation techniques such as JPEG 
compression. Our investigation aims to assess the resilience of robust hashing
methods against various forms of image manipulation, particularly in scenarios where 
images undergo alterations commonly encountered in digital forgery. Through 
experimental validation, our proposed fake image detection approach utilizing robust 
hashing demonstrates superior performance compared to existing state-of-the-art 
methods. We conduct comprehensive experiments using diverse datasets, including 
synthetic images generated with Generative Adversarial Networks (GANs), to 
evaluate the efficacy of our method across different image types and manipulation 
scenarios. Our findings underscore the effectiveness of robust hashing as a promising 
solution for detecting fake images in real-world applications, contributing to 
advancements in the field of image forensics and digital authentication.