Fig. 4: Comparison of UNSEG, Cellpose, and Mesmer on four example images from the GIT dataset.
From: UNSEG: unsupervised segmentation of cells and their nuclei in complex tissue samples
a Columns respectively correspond to appendix, esophagus, gallbladder, and small intestine tissue images. Rows show nucleus (white boundary) and cell (green boundary) segmentation results for the four examples using UNSEG, Cellpose and Mesmer, respectively. b The two rows, respectively, show nucleus and cell segmentation accuracy of UNSEG, Cellpose, and Mesmer. Accuracy is measured using number of segmented objects (see insets) and F1 score curves plotted as a function of IoU threshold between the segmented and annotated labels for nuclei and cells, respectively.
