Skip to main content
. 2022 Jan 19;11(2):156. doi: 10.3390/biology11020156

Figure 7.

Figure 7

The recognition result comparison. (a) Original image. (b) Simple threshold segmentation. (c) Comprehensive threshold segmentation. (d) Tiny YOLOv3. (e) Improved YOLOv3. (f) Gray level histogram. Simple threshold has the most noise and false-positive results due to the small difference in gray value between the colony and background; comprehensive threshold segmentation reduces noise interference but has more false-negative results; tiny YOLOv3 has a large improvement in accuracy but is less effective for small targets; improved YOLOv3 has optimal results.