Short survey: adaptive threshold methods used to segment immunonegative cells from simulated images of follicular lymphoma stained with 3,3'-Diaminobenzidine&Haematoxylin

Abstract

Abstract-We perform a short survey of image thresholding methods for very specific task, and assess their performance comparison. We analyse performance of adaptive thresholding methods concerning segmentation of immunonegative cells of follicular lymphoma tissue samples stained with 3,3'-Diaminobenzidine&Haematoxylin. We use artificial images based on experimental images that greatly simulates real samples and simplifies process of evaluation. We chose 8 methods of adaptive threshold segmentation, with different approach. They were applied to 6 different monochromatic images derived from original RGB images, by splitting layers, conversion to Lab colour space and colour deconvolution. Evaluation of the results was performed with basic statistical measures as sensitivity and specificity along with Jaccard's coefficient. We identify the thresholding algorithms with superior performance. Collected results will be used to design the new better method based on this approach.

Clustered nuclei splitting based on recurrent distance transform in digital pathology images

“…Then, they were split into images of separate punches using two designated programs [39,40]. Although the images are originally 3-channel Red-Green-Blue (RGB) images, we converted them to Hue-Saturaton-Value color-space and used only value layer for processing, similar to the previous research [41].…”