Efficient Deep Learning for Biometrics: Overview, Challenges and Trends in Ear of Frugal AI
1 st Karim Haroun, 2 nd Aya Zitouni †, 3 rd Aicha Zenakhri †, 4 th Meriem Amel Guessoum, 5 th Larbi Boubchir · 2026
Read Efficient Deep Learning for Biometrics: Overview, Challenges and Trends in Ear of Frugal AI by 1 st Karim Haroun, 2 nd Aya Zitouni †, 3 rd Aicha Zenakhri †, 4 th Meriem Amel Guessoum, 5 th Larbi Boubchir (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/arxiv-2602-08809.txt.
Description
Deep learning for body ID brings score gains but also heavy energy costs that hinder edge use and green costs. Karim Haroun and four co-authors survey lean deep learning for biometric uses, tackling training and use hurdles under the banner of frugal AI. Their arXiv paper sorts lean method groups and their trade-offs.
Extra leanness metrics, memory, compute, latency, and throughput, are covered too alongside a call for shared open metrics for fair match-ups. Next steps close the survey. Readers get a tour of how to shrink body-ID models for real-time, resource-limited use without losing sight of what leanness should mean.
Catalog record
| CREATOR | 1 st Karim Haroun, 2 nd Aya Zitouni †, 3 rd Aicha Zenakhri †, 4 th Meriem Amel Guessoum, 5 th Larbi Boubchir |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 30 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-08809