Publications

First Author

Vision-language models know more about agriculture than they show and rubric-grounded verifications close the gap
2026 · 0 citations VLMs possess latent agricultural knowledge that standard prompting fails to surface; rubric-grounded verification unlocks this knowledge and substantially closes the gap with supervised models.
Thermal image segmentation in weedy fields via synthetic RGB-trained models and GAN-based cross-modality alignment
2026 · 2 citations Combining synthetic training data with limited real annotations and GAN-based RGB-to-thermal translation significantly improves semantic segmentation of crops and weeds in complex field environments.
Are vision-language models ready to zero-shot replace supervised classification models in agriculture?
2025 · 7 citations Benchmarking diverse VLMs on 27 agricultural datasets reveals that current off-the-shelf models are not yet suitable as standalone diagnostic systems but can serve as assistive components with constrained interfaces and domain-aware evaluation.
AGILE: A diffusion-based attention-guided image and label translation for efficient cross-domain plant trait identification
2025 · 6 citations AGILE uses optimized text embeddings and attention-guided diffusion to achieve semantically consistent cross-domain image and label translation for plant trait identification.