From leaf to label: A robust automated workflow for stomata detection

Abstract

Plant leaf stomata are the gatekeepers of the atmosphere–plant interface and are essential building blocks of land surface models as they control transpiration and photosynthesis. Although more stomatal trait data are needed to significantly reduce the error in these model predictions, recording these traits is time‐consuming, and no standardized protocol is currently available. Some attempts were made to automate stomatal detection from photomicrographs; however, these approaches have the disadvantage of using classic image processing or targeting a narrow taxonomic entity which makes these technologies less robust and generalizable to other plant species. We propose an easy‐to‐use and adaptable workflow from leaf to label. A methodology for automatic stomata detection was developed using deep neural networks according to the state of the art and its applicability demonstrated across the phylogeny of the angiosperms. We used a patch‐based approach for training/tuning three different deep learning architectures. For training, we used 431 micrographs taken from leaf prints made according to the nail polish method from herbarium specimens of 19 species. The best‐performing architecture was tested on 595 images of 16 additional species spread across the angiosperm phylogeny. The nail polish method was successfully applied in 78% of the species sampled here. The VGG19 architecture slightly outperformed the basic shallow and deep architectures, with a confidence threshold equal to 0.7 resulting in an optimal trade‐off between precision and recall. Applying this threshold, the VGG19 architecture obtained an average F‐score of 0.87, 0.89, and 0.67 on the training, validation, and unseen test set, respectively. The average accuracy was very high (94%) for computed stomatal counts on unseen images of species used for training. The leaf‐to‐label pipeline is an easy‐to‐use workflow for researchers of different areas of expertise interested in detecting stomata more efficiently. The described methodology was based on multiple species and well‐established methods so that it can serve as a reference for future work.

HairNet: a deep learning model to score leaf hairiness, a key phenotype for cotton fibre yield, value and insect resistance

“…Recent studies have demonstrated that deep learning approaches can be used to quantify microscopic but important phenotypes such as the density and/or shape of stomata in a range of plant species [ 12 – 17 ]. Quantification of hairiness has been attempted in Arabidopsis thaliana , Soybean ( Glycine max ) and Spring Wheat ( Triticum aestivum ) [ 18 – 21 ].…”

GinJinn2: Object detection and segmentation for ecology and evolution

“…The stomata analysis serves as a basic example of instance segmentation. Despite several previous works on the automated examination of stomata (Toda et al, 2018;Fetter et al, 2019;Li et al, 2019;Carrasco et al, 2020;Casado-García et al, 2020;Meeus et al, 2020;Song et al, 2020), this contribution, to our knowledge, is the first trying to automatically segment whole stomata (represented by their guard cells) With the presented exemplary analyses, we hope to provide guidance for the application of GinJinn2 for automatic data collection and feature extraction. Despite GinJinn2's progress compared to its predecessor, there is still room for further improvements.…”

GinJinn2: Object detection and segmentation for ecology and evolution

“…The stomata analysis serves as a basic example of instance segmentation. Despite several previous works on the automated examination of stomata (Carrasco et al, 2020;Casado-García et al, 2020;Fetter et al, 2019;Li et al, 2019;Meeus et al, 2020;Song et al, 2020;Toda et al, 2018), this contribution, to our knowledge, is the first trying to automatically segment whole stomata (represented by their guard cells) using deep learning. With only 88 highly variable training images, our model achieved an AP of 51.32.…”