Publications/Speech Processing, Recognition & Audio Security
Journal · 2023
Speech Emotion Recognition for the Urdu Language
Nimra Zaheer, Obaid Ullah Ahmad, Mudassir Shabbir, Agha Ali Raza
Language Resources and Evaluation (LRE), 2023 · IF 1.358
Summary
Machines that recognize emotion in speech need large emotional speech datasets, which exist for only a handful of languages. We present SEMOUR+, the first scripted, emotion-tagged speech database for Urdu with diverse accents: 14 hours and 27,640 recordings by 24 native speakers, gender-balanced and phonetically balanced. Human raters agree strongly on the emotions conveyed. An ensemble model predicts the emotion of speakers it has never heard with 56% accuracy, a baseline for uses such as personalized assistants and gathering feedback from people who cannot use text-based technology.
Cite this paper
Nimra Zaheer, Obaid Ullah Ahmad, Mudassir Shabbir, Agha Ali Raza. Speech Emotion Recognition for the Urdu Language. Language Resources and Evaluation (LRE), 2023.
@article{zaheer2023speech,
author = {Nimra Zaheer and Obaid Ullah Ahmad and Mudassir Shabbir and Agha Ali Raza},
title = {{Speech Emotion Recognition for the Urdu Language}},
journal = {Language Resources and Evaluation},
year = {2023},
url = {https://aghaaliraza.com/publications/2022--LRE--Speech_Emotion_Recognition_for_the_Urdu_Language.pdf}
}
- Dissecting ASR Failures in Low-Resource South Asian LanguagesInterspeech 2026, Sydney, Australia
- WER We Stand: Benchmarking Urdu ASR ModelsCOLING 2025, Abu Dhabi, UAE
- Conversations in the Wild: Data Collection, Automatic Generation and EvaluationComputer Speech & Language, vol. 89, 2025
- Deepfake Defense: Constructing and Evaluating a Specialized Urdu Deepfake Audio DatasetFindings of ACL 2024, Bangkok, Thailand
- Using Self-Attention DNNs to Discover Phonemic Features for Audio Deep Fake DetectionASRU 2021, Cartagena, Colombia