Abstract
The ability to manipulate videos has been around for decades but a process that once would take time, money, and professionals, can now be created by anyone due to the rapid advancement of deepfake technology. Deepfakes use deep learning artificial intelligence to make fake digital content, typically in the form of swapping a person’s face in a video or image. This technology could easily threaten and manipulate individuals, corporations, and political organizations, so it is essential to find methods for detecting deepfakes. As the technology for creating deepfakes continues to improve, these manipulated videos are becoming increasingly undetectable. It is crucial to create methods to combat this problem. Previous research has been conducted on the various techniques to detect deepfakes, and though some models show promising results, many models struggle with reproducibility and practicality when exposed to real-world scenarios. Future work could consist of creating models without the tools used to generate deepfakes and the collected dataset in mind. Thus, the aim was to create a more general model that could be repeated on a variety of real data. To achieve this, a deepfake dataset was used to train models, and the results were analyzed. After comparing the strengths and limitations of previous models, we created simple, machine learning models that can accurately detect real-world deepfake images. Three methods, random forest, KNN, and SVM were utilized, and all achieved high accuracies compared to state-of-the-art models. Random forest had the best detection performance with accuracy results over 98%, followed by KNN and SVM. As deepfake technology continues to accelerate, it is essential to continue building models that can detect them because if not, it will be impossible to discern digital truth from reality.
Advisor
Rushit Dave
Committee Member
Rebecca Bates
Committee Member
Julie Wulfemeyer
Date of Degree
2023
Language
english
Document Type
Thesis
Program
Cognitive Science
Degree Program/Certificate
Cognitive Science with an emphasis in Computer Science
Degree
Bachelor of Science (BS)
Department
Computer Information Science
College
Science, Engineering and Technology
Recommended Citation
Conrad, D. (2023). A machine learning approach to deepfake detection [Bachelor of Science thesis, Minnesota State University, Mankato]. Cornerstone: A Collection of Scholarly and Creative Works for Minnesota State University, Mankato. https://cornerstone.lib.mnsu.edu/undergrad-theses-capstones-all/6/
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.
