Defakedev : Deep Fake Detector Integrated wth Virtual Universe |
Author(s): |
| Saurabh Satapathy , Odisha University of Technology and Research ; Ashis Kumar Mishra , Odisha University of Technology and Research ; Sumit Das Mohapatra , Odisha University of Technology and Research ; Omm Prakash Jena, Odisha University of Technology and Research |
Keywords: |
| Deep Learning, Deepfake Detection, Res-Next CNN, LSTM-based RNN, Face-Forensic++, Deepfake Detection Challenge, Celeb-DF, virtual universe, real-time detection, manipulated media, interactive platform |
Abstract |
|
Introducing a state-of-the-art deep learning method, our approach utilizes a Res-Next CNN for extracting frame-level features and an LSTM-based RNN for classification, effectively distinguishing between authentic and AI-generated fake videos. Trained on a diverse dataset comprising Face-Forensic++, Deepfake Detection Challenge, and Celeb-DF datasets, our method demonstrates robustness in identifying manipulated media. Furthermore, the integration of this deepfake detection model into a virtual universe allows for real-time detection of deepfakes, enriching user experiences within the simulated environment. This integration fosters greater awareness of the risks associated with manipulated media and offers an interactive platform for users to engage with and comprehend the implications of deepfake technology. |
Other Details |
|
Paper ID: IJSRDV12I10144 Published in: Volume : 12, Issue : 1 Publication Date: 01/04/2024 Page(s): 217-221 |
Article Preview |
|
|
|
|
