Detecting Fake Images Using Convolutional Neutral Networks - A Deep Learning Approach

Abstract

One popular generative model with many uses is the Generative Adversarial Network (GAN). Based on this unique concept, recent investigations have shown that it is feasible to produce high-quality fake face photos. The misuse of those fictitious faces in picture manipulation might lead to moral, ethical, and legal issues. To identify fake face images produced by the best method available at the moment, we first propose a Convolutional Neural Network (CNN) based method in this paper [20]. We also present experimental evidence demonstrating that the proposed method can achieve satisfactory results with an average accuracy over 99.4%. To further bolster the logic of our approach, we also offer comparison findings based on a few variations of the suggested CNN design, such as the high pass filter, the quantity of layer groups, and the activation function.