Abstract: Key traits, such as stomatal density, aperture, and pavement cell morphology, play a crucial role in plant physiology and ecology. However, manual quantification of these features is exceedingly labor-intensive and impedes research efficiency. While several automated analysis tools exist for examining stomata and pavement cells, they often fall short in providing both comprehensive functionality and user-friendliness. By integrating the latest detection and segmentation model, YOLOv11, with geometric algorithms, we developed a deep learning-powered tool, StomataQuant. It boasts an intuitive graphical user interface compatible with standard personal computers, enabling automated multi-task analysis and feature extraction for stomata detection, stomata and pores segmentation, and stomata and pavement cells segmentation. Its interactive editing interface offering manual correction significantly improved the efficiency of detection and data analysis. Systematic evaluations across diverse datasets demonstrate that StomataQuant exhibits exceptional concordance with manual measurements in stomata detection, stomata and pores segmentation, and stomata and pavement cells segmentation tasks. In practical applications, StomataQuant also yields conclusions consistent with manual measurements on both stomatal and pavement cell morphology. In the present study, we highlighted its powerful automated stomata and pavement cell detection with segmentation capabilities, and expect StomataQuant to significantly accelerate research advancements in plant physiology and ecology studies.

Key words: stomata, pavement cell, microscopy image analysis, applied deep learning, automated segmentation