Training a vision transformer on 2.5 million wheat images outperforms general-domain backbones across ten crop vision tasks.
https://www.kaggle.com/datasets/gauravduttakiit/cgiar - wheat-growth-stage-challenge
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
FoMo4Wheat: Toward reliable crop vision foundation models with globally curated data
Training a vision transformer on 2.5 million wheat images outperforms general-domain backbones across ten crop vision tasks.