Skip to main content
. 2017 Aug 18;7:8738. doi: 10.1038/s41598-017-09315-w

Figure 3.

Figure 3

Subset of Paired ROIs used to train the DL-CNN. Each ROI is 32 × 32 pixels. (a) ROIs that were labeled as being stage T0 after treatment. (b) ROIs that were labeled as being greater than stage T0 after treatment. A portion of ROIs in each class is zoomed in to illustrate the content of typical ROIs.