DeepFakeX: A Comprehensive Multimodal Deepfake Dataset for Research and Analysis

A multimodal deepfake dataset of 800 generated videos across four manipulation categories, in English and Urdu.

Data, 2026.

Note

Covers identity replacement, audio substitution, voice cloning and combined audiovisual alteration, with quality assessed by SSIM, MFCC and PCA. The inclusion of Urdu is deliberate: almost all existing deepfake datasets are English-only.

Reference

DeepFakeX: A Comprehensive Multimodal Deepfake Dataset for Research and Analysis
Data, 2026.
doi:10.3390/data11060141

The publisher record holds the author list.

Details

TypeDataset paper
AreasResponsible and generative AI
DOI10.3390/data11060141

Publisher record

More from the index

PreviousDINOv3-Powered Multi-Task Foundation Model for Quantitative Remote SensingAAAI
NextBridging Accuracy and Explainability: A SHAP-Enhanced CNN for Skin Cancer DiagnosisBook chapter

@layer base { [data-reveal] { opacity: 1 !important; transform: none !important; } #drawer[hidden] { display: block !important; } @media (min-width: 56.25rem) { #drawer[hidden] { display: none !important; } } .nav-toggle { display: none !important; } }