Deepfake Detection Analyzing Hybrid Dataset Utilizing CNN and SVM

View on arXiv ← Back to list

Authors: Jacob mallet, Laura Pryor, Rushit Dave, Mounika Vanamala

Published: 2023-01-27 01:00:39+00:00

AI Summary

This paper proposes a deepfake detection method using a hybrid approach combining Convolutional Neural Networks (CNN) and Support Vector Machines (SVM) on a dataset of 140,000 real and fake faces. The model achieved high accuracy in detecting deepfake images.

Abstract

Social media is currently being used by many individuals online as a major source of information. However, not all information shared online is true, even photos and videos can be doctored. Deepfakes have recently risen with the rise of technological advancement and have allowed nefarious online users to replace one face with a computer generated face of anyone they would like, including important political and cultural figures. Deepfakes are now a tool to be able to spread mass misinformation. There is now an immense need to create models that are able to detect deepfakes and keep them from being spread as seemingly real images or videos. In this paper, we propose a new deepfake detection schema using two popular machine learning algorithms.

Key findings

The CNN model outperformed the SVM model across all metrics, achieving an accuracy of 88.33%. Both models demonstrated high precision and recall, indicating good performance in correctly classifying both real and fake images. The limitations of model transferability and computational cost for real-world applications were acknowledged.

Approach

The authors utilize a hybrid approach, employing a CNN for feature extraction followed by an SVM for classification in one model and a CNN-only model for comparison. Data augmentation techniques were used to enhance the dataset.

Datasets

140k Real and Fake Faces dataset (70,000 real faces from Flickr-Faces-HQ and 70,000 deepfakes generated by StyleGAN)

Model(s)

Convolutional Neural Network (CNN) and Support Vector Machine (SVM)

Author countries

USA

← Previous