Application of convolutional neural networks towards nuclei segmentation in localization-based super-resolution fluorescence microscopy images

Table 1 Average test accuracy scores for Mask-RCNN trained on Kaggle dataset and tested on super-resolution imagery

Test set	Pre-processing	F1-Score	FN	Hausdorff
Colon tissue	512 × 512	0.181	0.819	14.07
	256 × 256	0.268	0.619	8.68
	256 × 256 Blur	0.262	0.635	8.03
	256 × 256 HEq	0.352	0.501	8.22
Cell line	512 × 512	0.073	0.924	12.83
	256 × 256	0.475	0.473	6.9
	256 × 256 Blur	0.555	0.268	5.92
	256 × 256 HEq	0.628	0.201	6.07

Average F1-Score, false negative percent (FN) and Hausdorff distance for a Mask R-CNN segmentation network model trained on the Kaggle dataset, and applied to both our super-resolution colon tissue and DNA labelled cell line datasets. The network was applied to our Colon Tissue and Cell Line image test sets (512 × 512 resolution), as well as to the downsized versions of each test set (256 × 256 resolution), and to Gaussian blurred (Blur) and histogram equalized (HEq) versions

ISSN: 1471-2105