Skip to main content
Figure 11 | BMC Medical Imaging

Figure 11

From: Automated detection of regions of interest for tissue microarray experiments: an image texture analysis

Figure 11

Identification of the normal-specific (shown in green), cancer-specific (shown in red), and non-specific (shown in yellow) appearance clusters on histology slides based on texture parameters computed using grayscale and color tissue segmentation algorithms. Relatively few image blocks are identified as normal specific (shown in green), while the cancer-specific image blocks conspicuously identify the tumors in the cancerous histology slides (shown in red). Overall, the delineations obtained by color tissue segmentation are more agreeable than those obtained using grayscale segmentation, as the latter misidentifies a few image blocks in normal histology slides as specific to cancer. Top two histology slides indicate normal tissue, whereas the bottom four have IDC, are shown at 1.25 times their actual size.

Back to article page