Fine-tuning SAM 3 on procedurally generated synthetic tomato-greenhouse images improves real-world fruit segmentation IoU from 52.7% to 65.6%.
Evaluating the efficacy of segment anything model for delineating agriculture and urban green spaces in multiresolution aerial and spaceborne remote sensing images
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Text-conditioned Segmentation for Tomato Phenotyping via Procedural Synthetic Data
Fine-tuning SAM 3 on procedurally generated synthetic tomato-greenhouse images improves real-world fruit segmentation IoU from 52.7% to 65.6%.