• Version
• Download 21
• File Size 322.48 KB
• File Count 1
• Create Date 30 October 2025
• Last Updated 30 October 2025

NoMask.AI – AI Vs Human Face Detection

Mrs.I.A.Jannathul Firthous1, Laksha.K2, Mawiyah.H3, Rithika.K.P4

1,2,3,4Information Technology, Sri Shakthi Institute of Engineering and Technology

Abstract - This document expressions the rapid advancements in artificial intelligence have enabled the creation of hyper-realistic synthetic faces that are nearly indistinguishable from genuine human faces. While this represents a significant technological milestone, it also poses critical challenges related to digital ethics, identity verification, and online security. NoMask.AI addresses this growing concern by providing an intelligent solution capable of distinguishing between real and AI-generated faces. The system employs a refined ResNet-18 convolutional neural network, trained on diverse datasets containing both authentic and synthetic facial images. Using OpenCV for face detection and preprocessing, and PyTorch for deep learning model inference, NoMask.AI performs probability-based classification to determine facial authenticity. A user-friendly Flask-based web interface, designed with HTML and CSS, enables seamless interaction and real-time detection. By integrating deep learning with practical usability, NoMask.AI enhances trust and security in digital environments, offering a robust defense against AI-generated identity fraud.

Key Words: Artificial Intelligence, Deepfake Detection, ResNet-18, Convolutional Neural Network, OpenCV, Digital Security

Download