VeriVoice-V1 Bilingual Speech Corpus
A bilingual Urdu–English speech corpus for deepfake detection and remote identity-verification research, released on IEEE DataPort.
Keywords
Overview
As Principal Investigator of the EPSRC project MIMER, I released VeriVoice-V1 (IEEE DataPort, 2026), built with industrial partner STech.ai for deepfake-resistant identity proofing.
The corpus supports research on bilingual Urdu–English speech, closing a gap in existing anti-spoofing benchmarks that are dominated by monolingual English material. It is one of three public datasets released under MIMER, alongside MSS and iRAW.
VeriVoice-V1 was the project's first public dataset deliverable and is intended for researchers working on remote identity verification, presentation-attack detection, and audio deepfake forensics.
Related Project
Released under the EPSRC-funded MIMER collaboration with UET Taxila and STech.ai.
View the MIMER project →