Fig. 3.

Overview of typical measurements performed on image data containing stomata—example of a representative dicot leaf. (A) Detection of stomata in images can be used for counting stomata or assessment of stomatal density. (B) Extraction of individual stoma can be used to calculate morphometry measurements including areas and dimensions. (C) Depending on the deep learning network used, bounding box methods may lead to incorrect morphometry measurements if the stomata are not orientated along the major axes. Note that this diagram has been slightly re-sized to emphasize the difference.