Fig. 1

Overview of DiSCount. a The DiSCount workflow. After image acquisition and processing, DiSCount infers the number of germinated and total number of Striga hermonthica seeds in each image (ca. 3 or 0.1 s per image on a CPU and a GPU, respectively) and tabulates the obtained results in a *.csv file. b Example of annotated Striga hermonthica seeds (green bounding boxes) and their radicle onsets (red bounding boxes). c Example of DiSCount’s visual output. The visual output is a *.png file displaying seed objects (purple rectangles) and radicle objects (i.e., germinated seeds; bright green rectangles). The debris present from impurities in the seed batch are mostly not recognized by DiSCount, which indicates the potential use of DiSCount with less clean Striga seed batches