Fig. 3.

Output from individual steps of the reconstruction pipeline from a clinical dataset of human bladder: (a) a representative, original WLC image, (b) point cloud from the structure-from-motion step before outlier removal, (c) mesh from the mesh-generation step, and (d) labeled texture (faces with the same color are labeled with the same input image) and (e–f) textured mesh from texture-generation steps. The green box shows a similar region between subfigures (d–f) indicating clear continuity of vessels despite the use of multiple input images to construct this region. The green box is approximately the size of a single WLC image. Black arrows in (a) and (f) indicate similar regions of the bladder.