Lip reading systems for Urdu alphabets in diverse environments
Amanullah Baloch, Mushtaq Ali, Lal Hussain, Ali Daud, Hussain Dawood · 2026
Read Lip reading systems for Urdu alphabets in diverse environments by Amanullah Baloch, Mushtaq Ali, Lal Hussain, Ali Daud, Hussain Dawood (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/PMC13527132.txt.
Description
Amanullah Baloch and colleagues build, in a 2026 study, lip-reading systems for Urdu alphabets, a capability that can help deaf users, noisy settings, and silent password entry. Urdu lacked a suitable dataset, and adapting an established model called LipNet proved difficult. The team created a new dataset of Urdu alphabet lip movements.
Three neural network designs were compared: a LipNet-based two-dimensional model, a hybrid model mixing two- and three-dimensional processing, and a plain three-dimensional baseline. All were helped by data augmentation and tested in varied environments on familiar and unseen speakers. The LipNet-based model reached about eighty-two percent accuracy on unknown data, leading on precision, recall, and F1 as well. The hybrid model generalized best among the others, reaching about sixty-nine percent on unfamiliar data.
Catalog record
| CREATOR | Amanullah Baloch, Mushtaq Ali, Lal Hussain, Ali Daud, Hussain Dawood |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 88886 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |