DTOFW: a lightweight tri-branch time-frequency fusion network for video-based recognition of diarrhea-related calf behavior
Wangli Hao, Qingqing Li, Lingling Li, Yakui Xue, Meng Han, Yanhong Liu, FuZhong Li · 2026
Read DTOFW: a lightweight tri-branch time-frequency fusion network for video-based recognition of diarrhea-related calf behavior by Wangli Hao, Qingqing Li, Lingling Li, Yakui Xue, Meng Han, Yanhong Liu, FuZhong Li (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/PMC13472929.txt.
Description
Hao and colleagues present a 2026 video analysis network that recognizes diarrhea-related behavior in calves, helping farms spot sick animals early. Their lightweight design, named DTOFW, reads video through a frozen vision backbone plus three branches that each capture a different signal pattern.
One branch models smooth motion over time, a second highlights repeating periodic patterns, and a third catches brief sudden changes linked to diarrhea. On the Jinnan calf behavior dataset, the system reached about 95 percent recognition accuracy. It beat the tested baseline models by up to about nine percentage points, while its trainable fusion head held only about 1.3 million parameters, a small size for deployment. The authors present the model as an efficient tool for fine-grained livestock health monitoring.
Catalog record
| CREATOR | Wangli Hao, Qingqing Li, Lingling Li, Yakui Xue, Meng Han, Yanhong Liu, FuZhong Li |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50 KB |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |