DeepBranchAI: a transferable 3D segmentation model for branching networks
Alexander V. Maltsev, Lisa M. Hartnell, Luigi Ferrucci · 2026
Read DeepBranchAI: a transferable 3D segmentation model for branching networks by Alexander V. Maltsev, Lisa M. Hartnell, Luigi Ferrucci (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-13623868.txt.
Description
Alexander Maltsev and co-authors present, in a 2026 study, a 3D segmentation model built for branching networks, the tree-like webs found in vessels, roots, and airways. Cutting such webs out of a volume scan is delicate: one wrong voxel can snap a branch or fake a new one. Slice-by-slice methods cannot guard the joins. Their model, DeepBranchAI, was trained on painstaking labels of a muscle cell's power-plant web, caught at nanometer detail.
Starting from that training beat starting from scratch on three outside tasks: liver veins, plant roots, and airways. Scores that reward unbroken branches rose in each case, most for the liver veins. A count of broken pieces fell sharply for the roots. One hard training ground, the authors argue, can teach a model to respect branches everywhere.
Catalog record
| CREATOR | Alexander V. Maltsev, Lisa M. Hartnell, Luigi Ferrucci |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 56 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |