Year: 2025
Type: article
Abstract: Disinformation and misinformation can be spread through fake images. Fake images can be employed to influence decision-making and manipulate public opinion. Fake image detection finds,use in a number of domains including law enforcement, national security, and social media.It can also be utilized in preventing the diffusion of misinformation and disinformation. The paper recommends a deep learning approach to the detection of forged images based on transfer learning. We utilize a pre-trained CNN weights and adjust them on a set of images used to fake or not, in order to produce a system which is a large improvement for the detection on our dataset and provides a 99% accuracy. We also gain insights into how various CNN structures and transfer learning methods are effective in detecting false images. Our work contributes to the creation of effective image forensics software, with real-world applications ranging from digital media verification to cyber security and policing.
Authors: Dr. Panguluri Vinodh Babu, Musunuri Naga Madhu, Galeeb Shaik, Kornipati Sravani, Mohammed Nayeemur Rahman
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4408973160