A Comprehensive Evaluation of Fake Face Recognition Scheme using Artificial Intelligence Oriented Learning Scheme
S. Subashree, Remya Rose S, A.S. Valarmathy, Princy Joseph, P. Jasmin Dennis, S. Ravi · 2024
Fake face recognition has emerged as a critical area of research due to the proliferation of synthetic media and its potential misuse in various domains. This study presents a thorough evaluation of a novel Fake Face Recognition Scheme utilizing Artificial Intelligence Oriented Learning Scheme (AIOLS) coupled with Generative Adversarial Network (GAN)-GoogleNet integration. The aim is to discern authentic facial images from synthetic or tampered ones with high accuracy. In this research, we employed a dataset comprising a diverse range of authentic and synthetic facial images. The proposed scheme incorporates the power of GANs for generating synthetic facial images and leverages GoogleNet, a state-of-the-art deep convolutional neural network, for discriminative feature extraction. The integration within the AIOLS framework facilitates efficient learning and adaptation to complex patterns present in fake facial images. The evaluation of our scheme was conducted through rigorous experimentation on the dataset, employing various performance metrics. Notably, the accuracy obtained was measured at an impressive 96.7%, signifying the effectiveness of our approach in distinguishing between real and fake facial images. Additionally, other metrics such as precision, recall, and F1-score were also computed to provide a comprehensive assessment of the scheme's performance.