Table 2. Results of the YOLO v3 network for ear tip detection on images of wheat plants sourced from internet databases (Fig. 9).

Left, left middle, right middle, and right correspond to the position of the image on the top row of Figure 9. The images from the bottom row of Figure 9 are not included because no ears were found, which is the expected result of the network as ears with awns are classified as incorrect. The All images row includes all images that were used. “Detected” refers to the number of found ears, whereas “undetected” are the total number of ears missed. “Accuracy” is the percent of detected ears.

Image	Detected	Undetected	Actual Ears	Accuracy (%)
Left	4	3	7	57
Left-middle	11	2	13	85
Right-middle	16	4	20	80
Right	8	2	10	80
All images	229	47	276	83