Skip to main content
Fig. 6 | BMC Biology

Fig. 6

From: Extracting multiple surfaces from 3D microscopy images in complex biological tissues with the Zellige software tool

Fig. 6

Organoid specimen. A,B Volume rendering (A) and orthogonal sections (B) of a confocal 3D image of a (half of) inner ear organoid, which has been fixed and stained with phalloidin to reveal F-actin. The dataset (of dimensions 520 × 465 × 35 voxels) includes two dome-shaped epithelial surfaces of interest, forming the apical (inward) and basal (outward) sides of the organoid. C 3D representations of the height map extracted by Zellige (in green) and the GT height map (in blue), of the epithelium surface. D Color-coded error maps of the reconstructed height maps for the apical (left) and basal (right) epithelial surfaces of the organoid. The surfaces of interest are reconstructed with an error of < 2 pixels over a large majority (96% and 93% for the apical and basal surfaces, respectively) of pixels, as well as on average (RMSE ~ 0.8 and 1.1 for the apical and the basal surfaces, respectively). E Projections localized to the GT height maps of the epithelium surface (panels on the left), and the height maps extracted by Zellige (panels on the right). F Using the DeProj tool (Herbert et al. [12]) to generate cell morphology measurements corrected for the projection factor associated with the local surface slope. Cells of the surface S2, which represents the epithelial apical surface, were segmented using the TissueAnalyzer tool (Etournay et al. [3]). The height map obtained from Zellige and the segmentation mask were used to calculate geometrical parameters of the surface and its constituent cells. From left to right: the local slope, the local mean curvature, the projection error on cell area due to the local slope, and the corrected cell area in a 3D representation (axis ratio 1:1:3), for the reconstructed surface S2. Scale bar 100 μm

Back to article page