Figure 3: 3D image data reconstruction.

Each CT image scan consists of multiple slices. In the 3D image reconstructions, a block was laid over the eyes for visual de-identification, similar to so-called censor bars in the eye area of standard 2D photos. (a) Original 3D image reconstruction based on multiple slices. (b) 3D image reconstruction with the slice-based segmentation mask (3D ground truth mask). Note: Fig. 3 was previously released in part17.