New framework uses synthetic data to improve tomato plant segmentation

Researchers have developed a new framework for segmenting tomato plants in greenhouses, addressing the challenge of limited annotated training data. This approach combines procedural synthetic data generation with fine-tuning of the Segment Anything Model 3 (SAM 3). By modeling a commercial cherry tomato greenhouse, they created a large synthetic dataset that was used to specialize SAM 3's text-conditioned segmentation capabilities for crop organs. The fine-tuned model demonstrated significantly improved segmentation performance and confidence when evaluated on real-world greenhouse datasets. AI

IMPACT Enhances AI's applicability in specialized agricultural environments by improving segmentation accuracy.

RANK_REASON Academic paper detailing a new method for computer vision tasks. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• Gotit.pub
• Hugging Face
• Mikolaj Cieslak
• ScienceCast
• Segment Anything Model 3
• tomato

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →