Paper Contents
Abstract
The rapid advancement of deepfake technology has raised critical concerns regarding the authenticity of digital visual content. This study proposes a deep learning-based system for the detection of deepfake images utilizing a Convolutional Neural Network (CNN) framework. The model is trained on a diverse dataset containing both real and synthetically generated images, processed through systematic preprocessing steps such as normalization, resizing to 224224 pixels, RGB color space conversion, and data augmentation. The CNN architecture extracts salient features and employs sigmoid activation for binary classification, producing confidence scores to indicate prediction reliability. For real-time usability, the model is integrated into a Flask-based web application that allows users to upload and verify image authenticity instantly. The system's effectiveness is validated using standard performance metrics including accuracy, precision, recall, and F1-score. This research offers a scalable and accessible solution to mitigate the risks posed by manipulated media, aiding digital forensics and enhancing trust in visual content.
Copyright
Copyright © 2025 RAMYA B N. This is an open access article distributed under the Creative Commons Attribution License.
