Figure 2 | Scientific Reports

Figure 2

From: Comparison between a deep-learning and a pixel-based approach for the automated quantification of HIV target cells in foreskin tissue

Figure 2

Automated segmentation of HIV target cells in foreskin tissue using pixel-based and StarDist models. (A,E) Representative image of (A) nuclei staining and (E) composite image of staining for CD3, CD4, CCR5, and nuclei were generated from a cropped image of a full tissue section scan of foreskin tissue imaged at 200 × total magnification. Visual representation of (B–D) counting of all positively stained nuclei and (F–H) CD3 + CD4 + CCR5 + cells in the representative image. Methods of cell segmentation presented include: (B,F) manual counting of cells, (C,G) automated counting using a pixel-based model, and (D,H) automated segmentation using a custom StarDist deep-learning model.

Back to article page