Drone‐Assisted Image Forgery Detection Using Generative Adversarial Net‐Based Module
Swathi Gowroju, Shilpa Choudhary, M. Rishitha, S. Tejaswi, Lankala Shashank Reddy, M. Sujith Reddy · 2024
In today's digital world, image forgery detection in drone-captured images is challenging due to noise, motion blur, and other factors unique to aerial photography. The research community is actively seeking solutions to this problem. Biometric systems are becoming increasingly important due to each individual's unique identity. However, some people modify their physical appearance to avoid detection by these systems, highlighting the need for advanced forgery detection techniques. We propose an LBPNet, an LBP-based machine learning convolutional neural network that can detect fake face photos to address this issue. The proposed System compares LBPNet and NLBPNet, as it relies on feature extraction using the LBP algorithm. Additionally, our suggested paired learning technique enables forged feature learning, allowing the detection module to identify falsified images generated by a new GAN, even if it was not included in the training phase. Using a drone-based system can capture high-resolution images and videos from different angles and perspectives, further enhancing the efficiency and speed of the detection process. Our proposed approach can significantly improve image forgery detection, making it a valuable tool for various industries, including journalism, law enforcement, and national security.