Investigating SOTA Deepfake Detection frameworks in content-driven multimodal deepfake content

Gupte, Abhishek Abhay (2025) Investigating SOTA Deepfake Detection frameworks in content-driven multimodal deepfake content. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (4MB) | Preview
Preview	PDF (Configuration Manual) Download (6MB) | Preview

Abstract

Deepfake generation technology has evolved over the past several years in unprecedented ways. A majority of such media are either entirely fake or entirely real & are reasonably easy to identify using current state-of-the-art binary-level classification deepfake detection systems. However, in recent years a new form of deepfake temporal forgery dataset, called the LAV-DF has emerged. This consists of deepfake videos where only a few temporal regions have been manipulated to change the sentiment/meaning of the entire video has emerged. Current traditional deepfake detection architectures need severe modification in order to keep up with the accuracy of detecting temporal localised datasets, which require a compromise in terms of environmental efficiency. This work has focused on implementing a version of a popular SOTA multimodal deepfake detection architecture called the NPVForensics - trained on the LAV-DF, evaluating weaknesses in the architecture and point to future directions. The results have shown the SOTA architecture under constrained resources has not performed well on the dataset. The future directions propose implementation and fine-tuning approaches to mitigate difficulties.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Sahni, Anu UNSPECIFIED
Subjects:	Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence Z Bibliography. Library Science. Information Resources > ZA Information resources > ZA4150 Computer Network Resources > The Internet > World Wide Web > Websites > Online social networks T Technology > TK Electrical engineering. Electronics. Nuclear engineering > Telecommunications > The Internet > World Wide Web > Websites > Online social networks
Divisions:	School of Computing > Master of Science in Artificial Intelligence
Depositing User:	Ciara O'Brien
Date Deposited:	02 Sep 2026 09:26
Last Modified:	02 Sep 2026 09:26
URI:	https://norma.ncirl.ie/id/eprint/9757

Actions (login required)

View Item