Skip to main content
. Author manuscript; available in PMC: 2022 Feb 1.
Published in final edited form as: Med Image Anal. 2020 Nov 11;68:101892. doi: 10.1016/j.media.2020.101892

Figure 8.

Figure 8.

Example results of three deep-learning detection-based cell counting methods (Mask R-CNN, U-Net, and StructRegNet). Panels (a) and (b) show the prediction results on the bacterial and bone marrow datasets, respectively. The green cycles and red dots in each image represent the ground truth annotations and the detected cell centroids, respectively.