Saved in:
Bibliographic Details
Main Authors: Deoli, Pankaj, Tej, Atef, Ashri, Anmol, JS, Anandatirtha, Berns, Karsten
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2604.13722
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911595260542976
author Deoli, Pankaj
Tej, Atef
Ashri, Anmol
JS, Anandatirtha
Berns, Karsten
author_facet Deoli, Pankaj
Tej, Atef
Ashri, Anmol
JS, Anandatirtha
Berns, Karsten
contents We address the challenge of synthetic-to-real transfer in forestry perception where real data have only coarse Tree labels while synthetic data provide fine-grained trunk/crown annotations. We introduce MGTD, a mixed-granularity dataset with 53k synthetic and 3.6k real images, and a four-stage protocol isolating domain shift and granularity mismatch. Our core contribution is granularity-aware distillation, which transfers structural priors from fine-grained synthetic teachers to a coarse-label student via logit-space merging and mask unification. Experiments show consistent mask AP gains, especially for small/distant trees, establishing a testbed for Sim-Real transfer under label granularity constraints.
format Preprint
id arxiv_https___arxiv_org_abs_2604_13722
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Granularity-Aware Transfer for Tree Instance Segmentation in Synthetic and Real Forests
Deoli, Pankaj
Tej, Atef
Ashri, Anmol
JS, Anandatirtha
Berns, Karsten
Computer Vision and Pattern Recognition
We address the challenge of synthetic-to-real transfer in forestry perception where real data have only coarse Tree labels while synthetic data provide fine-grained trunk/crown annotations. We introduce MGTD, a mixed-granularity dataset with 53k synthetic and 3.6k real images, and a four-stage protocol isolating domain shift and granularity mismatch. Our core contribution is granularity-aware distillation, which transfers structural priors from fine-grained synthetic teachers to a coarse-label student via logit-space merging and mask unification. Experiments show consistent mask AP gains, especially for small/distant trees, establishing a testbed for Sim-Real transfer under label granularity constraints.
title Granularity-Aware Transfer for Tree Instance Segmentation in Synthetic and Real Forests
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.13722