Quantifying the Frontal-to-Ceiling Domain Gap for YOLO-Based Hand Gesture Recognition in Smart Homes
Ufuk Beşenk, Sarp Ege Nayim, Ömür Öcal, Mehmet Öztemel, Ahmet Özkurt, Mustafa Alper Selver · 2026
Read Quantifying the Frontal-to-Ceiling Domain Gap for YOLO-Based Hand Gesture Recognition in Smart Homes by Ufuk Beşenk, Sarp Ege Nayim, Ömür Öcal, Mehmet Öztemel, Ahmet Özkurt, Mustafa Alper Selver (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/pmc-13611033.txt.
Description
Ufuk Beşenk and five co-authors quantify the domain gap between frontal and ceiling camera views for hand gesture recognition in smart homes. Gesture systems are built for frontal views, but home cameras often sit on ceilings. That viewpoint shift could quietly degrade recognition where the systems are actually deployed.
The team builds CeilGest, a ceiling view dataset with 18 gesture classes, 156,282 annotated frames, and 68 participants in real homes. Detectors trained on a large frontal dataset are tested against a model trained in domain on identical footage. The frontal model misclassifies about 24 times more frames, 486 against 20 in 27,000 frames. The damage is class dependent, while a lightweight model runs the full pipeline on small edge devices.
Catalog record
| CREATOR | Ufuk Beşenk, Sarp Ege Nayim, Ömür Öcal, Mehmet Öztemel, Ahmet Özkurt, Mustafa Alper Selver |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 74 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |