Fauth AI – A Multimodal Deepfake Detection using Deep Learning

Abarna, N and Abinaya, G and DEEPA, R (2025) Fauth AI – A Multimodal Deepfake Detection using Deep Learning. In: FAUTH AI - A Multimodal Deepfake Detection using Deep Learning, 25.10.2025, CHENNAI.

[thumbnail of Gogan college conference.pdf] Text
Gogan college conference.pdf

Download (78MB)

Abstract

FauthAI is an end-to-end framework for deepfake detection integrating CNNs for local artifact
detection, RNNs/LSTMs for temporal inconsistency analysis, and Vision Transformers for
global feature extraction. The system also employs frequency domain and audio-based analysis
for spectral anomalies and voice signature evaluation. Multimodal fusion of audio-visual data
and ensemble learning enhances detection accuracy across datasets such as DFDC, Face
Forensics++, Celeb-DF v2, and ASVspoof. The framework offers a scalable and robust solution
for digital content authentication, media forensics, and misinformation management.

Item Type: Conference or Workshop Item (Paper)
Subjects: Computer Science Engineering > Deep Learning
Domains: Computer Science Engineering
Depositing User: Mr IR Admin
Date Deposited: 16 May 2026 11:13
Last Modified: 16 May 2026 11:34
URI: https://ir.vistas.ac.in/id/eprint/19870

Actions (login required)

View Item
View Item