Multilevel neural networks with dual-stage feature fusion for human activity recognition
Abeer FathAllah Brery \corrauth, Ascensión Gallardo-Antolín, Israel Gonzalez-Carrasco and Mahmoud Fakhry \corrauth · 2026
Read Multilevel neural networks with dual-stage feature fusion for human activity recognition by Abeer FathAllah Brery \corrauth, Ascensión Gallardo-Antolín, Israel Gonzalez-Carrasco and Mahmoud Fakhry \corrauth (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-2604-16577.txt.
Description
Brery and colleagues explore multilevel networks with dual-stage feature fusion for activity recognition. Combining convolutional, memory and hybrid networks exploits matching strengths, but layout matters. Their two-level design uses late fusion of first-level outputs plus intermediate fusion across both levels. Testing 15 architectures on two public benchmarks, both fusions beat late fusion alone. The best configuration also beat baseline models, proving the approach for activity recognition.
Catalog record
| CREATOR | Abeer FathAllah Brery \corrauth, Ascensión Gallardo-Antolín, Israel Gonzalez-Carrasco and Mahmoud Fakhry \corrauth |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33026 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.16577