Tools
Tools
Sharma, Hitesh Kumar, Khan, Soumya Suvra, Choudhury, Tanupriya and Khurana, Madhu 
ORCID: 0000-0003-3976-1256
(2023)
CNN-Based Model for Deepfake Video and Image Identification Using GAN.
In:
Proceedings of Fourth International Conference on Computer and Communication Technologies.
Springer, Singapore, pp. 481-489.
ISBN 9789811985621
Abstract
Deepfakes are the new age tools that automate the syntheses and detection of computer altered videos through GANs. Studies and researches are being done to detect and study the impact of deepfakes on social media and on human lives. In this paper, we will research about the DF technologies such as MTCNN and ResNext-v1 classification models to artificially automate the tasks of deepflakes detection by using datasets from varied sources and having different diversities of people. We also portray another deep learning-based technique that can successfully recognize AI-created counterfeit recordings from genuine recordings. It is inconceivably critical to foster innovation that can spot fakes, so the DF can be recognized and kept from spreading over the Web. Our strategy identifies by looking at the facial zones and their encompassing pixels by parting the video into outlines and separating the highlights with a ResNext-v1 CNN and utilizing the MTCNN catch the transient irregularities between frames presented by GANs during the remaking of the pixels. Our aim is to make an audio-less deepfakes detection system using ML and DL techniques to curb the spread of misinformation. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
| Item Type: | Book Section |
|---|---|
| Article Type: | Article |
| Uncontrolled Keywords: | Artificial intelligence; Deep learning; Deepfakes; GANs; Machine learning; MTCNN; ResNext-v1 |
| Subjects: | Q Science > QA Mathematics > QA76 Computer software |
| Divisions: | Schools and Research Institutes > School of Business, Computing and Social Sciences |
| Depositing User: | Susan Turner |
| Date Deposited: | 05 May 2023 11:04 |
| Last Modified: | 31 Oct 2023 12:43 |
| URI: | https://eprints.glos.ac.uk/id/eprint/12696 |
University Staff: Request a correction | Repository Editors: Update this record
