A prototype-augmented few-shot segment anything model for field phenotyping of lettuce germplasm from UAV imagery

Tian Xia, Jiahui Qi, Songtao Ban, Mengen Yuan, Linyi Li, Dong Hu, Jifeng Ning, Guotao Huo, Hui Yang, Shuping He, Guojun Ge, Shiwei Wei, Minglu Tian · 2026

Read A prototype-augmented few-shot segment anything model for field phenotyping of lettuce germplasm from UAV imagery by Tian Xia, Jiahui Qi, Songtao Ban, Mengen Yuan, Linyi Li, Dong Hu, Jifeng Ning, Guotao Huo, Hui Yang, Shuping He, Guojun Ge, Shiwei Wei, Minglu Tian (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13571494.txt.

Description

Tian Xia and co-authors adapt a segment-anything model for lettuce trait mapping from drone images. Breeders must score germplasm lines fast and accurately in the field. Hand labeling of plant images is slow and costly.

General segmentation models lack crop-specific visual knowledge. Few-shot learning offers accuracy with very little labeled data. Their model adds prototype features to guide its prompts. A gated attention part sharpens focus across space and channels. The target is precise outlines of varied lettuce lines. Field scenes mix light, soil, and plant shapes heavily. Better outlines improve trait extraction for screening. The study supports drone phenotyping for breeding programs with large trial fields.

Catalog record

CREATORTian Xia, Jiahui Qi, Songtao Ban, Mengen Yuan, Linyi Li, Dong Hu, Jifeng Ning, Guotao Huo, Hui Yang, Shuping He, Guojun Ge, Shiwei Wei, Minglu Tian
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE60 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC13571494

Browse the full library · About the Tower of Babel