Sparse Representation Learning for Vessels
Chinmay Prabhakar, Bastian Wittmann, Paul Büschl, Hongwei Bran Li, Bjoern Menze, Suprosanna Shit ⋆ · 2026
Read Sparse Representation Learning for Vessels by Chinmay Prabhakar, Bastian Wittmann, Paul Büschl, Hongwei Bran Li, Bjoern Menze, Suprosanna Shit ⋆ (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-2605-01382.txt.
Description
Chinmay Prabhakar and five co-authors present a compact way to store and study whole networks of blood vessels at fine detail. Reading a full organ's vessels at clinical resolution is costly, and older methods cut the network into small patches or thin tree sketches.
Their model, VAEsselSparse, uses sparse math that skips empty space, plus attention steps, to shrink a 3D vessel map by a factor of eight in each direction. It rebuilds the input better than dense rivals and earlier methods. The small code it produces still carries signs doctors use, such as bulges and narrowings in vessel walls, and which variant of the brain's arterial circle is present. The codes can also seed models that grow new, realistic vessel networks.
Catalog record
| CREATOR | Chinmay Prabhakar, Bastian Wittmann, Paul Büschl, Hongwei Bran Li, Bjoern Menze, Suprosanna Shit ⋆ |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.01382