A two-stream network with global-local feature fusion for bone age assessment
Qiong Lou, Han Yang, Fang Lu · 2026
Read A two-stream network with global-local feature fusion for bone age assessment by Qiong Lou, Han Yang, Fang Lu (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-13396481.txt.
Description
Qiong Lou and co-authors present a deep learning system that reads a child's bone age from hand X-rays. Doctors use bone age to judge growth and maturity, but reading films by hand is slow. Their model, BoNet+, splits the task into two streams, one for the whole-hand view and one for fine bone detail.
A transformer module in the global stream weighs links across the full image. A special convolution module in the local stream builds attention maps at several scales to catch small detail. The two streams merge and pass through an Inception-V3 network. On two public test sets, the mean error was about four months and about six months. The authors say the tool can cut clinical workload.
Catalog record
| CREATOR | Qiong Lou, Han Yang, Fang Lu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 67 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |