StomataSeg: Semi-supervised instance segmentation for sorghum stomatal components
Computers and Electronics in Agriculture · 2026-11 · 254 · 112442
DOI: 10.1016/j.compag.2026.112442 ↗Kurz gefasst
Fachartikel in der Zeitschrift Computers and Electronics in Agriculture, 2026-11. Verfasst von Zhongtian Huang und 10 weiteren Autor:innen. Die Arbeit nennt 80 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Australian Plant Phenomics Network, Australian Research Council, National Collaborative Research Infrastructure Strategy.
Abstract (Originalsprache)
Sorghum is a globally important cereal grown widely in water-limited and stress-prone regions. Its strong drought tolerance makes it a priority crop for climate-resilient agriculture. Improving water-use efficiency in sorghum requires precise characterization of stomatal traits, as stomata control of gas exchange, transpiration and photosynthesis have a major influence on crop performance. Automated analysis of sorghum stomata is difficult because the stomata are small (often less than 40 μ m in length in grasses such as sorghum) and vary in shape across genotypes and leaf surfaces. Automated segmentation contributes to high-throughput stomatal phenotyping, yet current methods still face challenges related to nested small structures and annotation bottlenecks. In this paper, we propose a unified semi-supervised instance segmentation framework for multi-class segmentation of three nested sorghum stomatal components. We collect and annotate a sorghum leaf imagery dataset containing 11,060 human-annotated patches, covering the three stomatal components (pore, guard cell and complex area) across multiple genotypes and leaf surfaces. To improve the detection of tiny structures, we split high-resolution microscopy images into overlapping small patches. We then apply a pseudo-labelling strategy with class-specific confidence filtering and expert verification to unannotated images, producing an additional 56,428 pseudo-labelled patches. Patch-based preprocessing increases the top semantic segmentation mIoU from 65.93% to 70.35% and the top instance segmentation AP from 28.30% to 46.10%. Subsequent pseudo-label retraining further increases overall AP from 46.10% to 49.20% for Mask R-CNN with ConvNeXt-V2. These results demonstrate that combining patch-based preprocessing with semi-supervised learning significantly improves the segmentation of fine stomatal structures. The proposed framework supports scalable extraction of stomatal traits and facilitates broader adoption of AI-driven phenotyping in crop science.
Förderung
- Australian Plant Phenomics Network
- Australian Research Council
- National Collaborative Research Infrastructure Strategy
- German Research Foundation
Zitieren
APA
Huang, Z., Chen, Z., Huang, Z., Yu, X., Smith, D., Purushothama, C., Oosterom, E. V., Wu, A., Salter, W., Li, Y., & Chapman, S. (2026). StomataSeg: Semi-supervised instance segmentation for sorghum stomatal components. Computers and Electronics in Agriculture, 254, 112442. https://doi.org/10.1016/j.compag.2026.112442
Vancouver
Huang Z, Chen Z, Huang Z, Yu X, Smith D, Purushothama C, et al. StomataSeg: Semi-supervised instance segmentation for sorghum stomatal components. Computers and Electronics in Agriculture. 2026;254:112442. doi:10.1016/j.compag.2026.112442
BibTeX
@article{Huang2026,
author = {Huang, Zhongtian and Chen, Zhi and Huang, Zi and Yu, Xin and Smith, Daniel and Purushothama, Chaitanya and Oosterom, Erik Van and Wu, Alex and Salter, William and Li, Yan and Chapman, Scott},
title = {StomataSeg: Semi-supervised instance segmentation for sorghum stomatal components},
journal = {Computers and Electronics in Agriculture},
year = {2026},
volume = {254},
pages = {112442},
doi = {10.1016/j.compag.2026.112442}
}
EUVIMED