A Stomata Classification and Detection System in Microscope Images of Maize Cultivars

Abstract

Stomata are morphological structures of plants that have been receiving constant attention. These pores are responsible for the interaction between the internal plant system and the environment, working on different processes such as photosynthesis process and transpiration stream. As evaluated before, understanding the pore mechanism play a key role to explore the evolution and behavior of plants. Although the study of stomata in dicots species of plants have advanced, there is little information about stomata of cereal grasses. In addition, automated detection of these structures have been presented on the literature, but some gaps are still uncovered. This fact is motivated by high morphological variation of stomata and the presence of noise from the image acquisition step. Herein, we propose a new methodology of an automatic stomata classification and detection system in microscope images for maize cultivars.In our experiments, we have achieved an approximated accuracy of 97.1% in the identification of stomata regions using classifiers based on deep learning features. 1According to Willmer and Fricker [1], from all points of view, the stomata have received more 2 constant attention probably than any other single vegetative structure in the plant. Regulating 3 gas exchange between the plant and the environment[2], these structures are small pores on the 4 surfaces of leaves, stems and parts of angiosperm flowers and fruits [3, 4], formed by a pair of 5 specialized epidermal cells (guarder cells), which are found in the surface of aerial parts of most 6 higher plants [1]. Due to the controlling of the exchange of water vapour and CO 2 between the 7 interior of the leaf and the atmosphere [3]; the photosynthesis, the transpiration stream, the 8 nutrition and the metabolism of land plants are in different ways related to the opening and 9 closing movements of the stomata [4, 1]. Furthermore, Hetherington and Woodward point that 10 the acquisition of stomata and an impervious leaf cuticle are considered to be key elements in 11 the evolution of advanced terrestrial plants, allowing the plant to inhabit a range of different, 12 often fluctuating environments but still control water content [3].13 The stomatal movements distinguish this structure from other pores found in plant organs, of liverworts [1]. The control of stomatal aperture requires the coordinated control of multiple 16 cellular processes [3] and its morphogenesis is affected by several environmental stimuli, such 17 as relative humidity, temperature, concentration of atmospheric carbon dioxide, light intensity, 18 and endogenous plant hormones [2, 3, 1]. Global warming for example could increase leaf 19 transpiration and soil evaporation, and as consequence leaf stomata movements can control 20 plant water loss and carbon gain under this water stress condition [5]. Stomatal aperture 21 might also represent an initial response to both plants and human pathogenic bacteria [2]. In 22 plants, it has been reported that microscopic surface openings serve as...

Optical topometry and machine learning to rapidly phenotype stomatal patterning traits for maize QTL mapping

“…There have been many attempts to address the phenotyping bottleneck for stomatal patterning through computer-aided image analysis. Classical image processing methods ( Omasa and Onoe, 1984 ; Liu et al, 2016 ; Duarte et al, 2017 ) and machine learning models have been applied ( Vialet-Chabrand and Brendel, 2014 ; Higaki et al, 2015 ; Jayakody et al, 2017 ; Saponaro et al, 2017 ; Dittberner et al, 2018 ; Toda et al, 2018 ; Aono et al, 2019 ; Bhugra et al, 2019 ; Fetter et al, 2019 ; Li et al, 2019 ; Sakoda et al, 2019 ). Although a number of these methods have been demonstrated to work within constrained image sets, none of them have been widely adopted, even within a single species.…”

Automated stomata detection in oil palm with convolutional neural network

“…Though the precision values acquired from these studies were very high, the recall values were lower by comparison. Compared to these methods that use specific feature engineering methods, a more recent methodology of automating stomata detection using multiple feature extraction techniques and learning methods was introduced in maize 34 , achieving 97.1% detection accuracy (measured in precision). This study also highlighted the use of deep learning features for stomata detection application.…”

A Deep Learning-Based Method for Automatic Assessment of Stomatal Index in Wheat Microscopic Images of Leaf Epidermis

“…Stomata in the leaf epidermis are bounded by the bean-or dumbbell-shaped guard cells with fixed shapes, and in some species but not all, they are surrounded by one-to-many subsidiary cells (Boetsch et al, 1996). At present, various image analysis tools have been developed for detecting, counting (Aono et al, 2019;Fetter et al, 2019), and measuring stomatal aperture (Omasa and Onoe, 1984;Li et al, 2019) as well as assessing stomatal density (Vialet-Chabrand and Brendel, 2014). However, to the best of our knowledge, there is no pipeline designed for the stomatal index measurement, possibly due to the difficulty in epidermal cell detection.…”