Temporal Feature Weaving for Neonatal Echocardiographic Viewpoint Video Classification
Satchel French, Faith Zhu, Amish Jain, Naimul Khan · 2025
Read Temporal Feature Weaving for Neonatal Echocardiographic Viewpoint Video Classification by Satchel French, Faith Zhu, Amish Jain, Naimul Khan (2025) 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-2501-03967.txt.
Description
Satchel French and co-authors classify the viewing angle of newborn heart ultrasounds from video rather than single frames, in a 2025 paper. Knowing which standard view an echocardiogram shows is a first step in screening, and clinics without expert sonographers could use automated help. A still image discards the motion that human experts unconsciously use.
Their network pairs a convolutional front end with a recurrent unit and a new temporal feature weaving step that mixes information across frames. Using only four consecutive frames, it beat image-only classification by 4.33 percent in accuracy, at little extra computing cost. The team also released a professionally annotated neonatal dataset with sixteen viewpoints to spur further work in this underserved area.
Catalog record
| CREATOR | Satchel French, Faith Zhu, Amish Jain, Naimul Khan |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 17 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-03967