Granularity-Aware Transfer for Tree Instance Segmentation in Synthetic and Real Forests.
Pankaj Deoli, Atef Tej, Anmol Ashri, Anandatirtha JS, Karsten Berns · 2026
Read Granularity-Aware Transfer for Tree Instance Segmentation in Synthetic and Real Forests. by Pankaj Deoli, Atef Tej, Anmol Ashri, Anandatirtha JS, Karsten Berns (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-2604-13722.txt.
Description
Deoli and four colleagues tackle tree segmentation when real forest labels are coarse but synthetic data label trunks and crowns in fine detail. Their MGTD dataset pairs 53,000 synthetic images with 3,600 real ones, and a four-stage protocol separates domain shift from granularity mismatch. A student trained on coarse labels learns structure from synthetic teachers with fine labels, via merged logits and unified masks. Tests show clear gains, especially for small and distant trees.
Catalog record
| CREATOR | Pankaj Deoli, Atef Tej, Anmol Ashri, Anandatirtha JS, Karsten Berns |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 73027 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.13722