Fig. 5.

Registration of a 3D multichannel video using Fast4DReg. A 3D multichannel video of cancer cells migrating inside the lung vasculature was corrected using Fast4DReg. (A) 3D surface rendering of a selected time point (see also Movie 4) created using Arivis Vision4D. The lung vasculature is shown in grey, and the cancer cell (AsPC1) is in green. The red rectangle indicates the clipping plane used to render the image. (B) Schematic illustrating the pipeline used to correct a multichannel 3D video using Fast4DReg. For this dataset, the drift was first estimated using the vasculature images (channel 1), and the resulting drift table was then applied to the cancer cell images (channel 2). (C–E) Three consecutive frames of the vasculature (C) and cancer cell images (D) were pseudo-colored blue, green and red, and merged. White indicates structural overlaps between the three frames. Line profiles along the dotted white lines to further study the overlap between frames were drawn as shown. (C,E) Z-projections are displayed to visualize the lateral misalignments corrected by Fast4DReg. Scale bars: 10 μm. (D,F) Y-projections are displayed to visualize the axial misalignment corrected by Fast4DReg. Scale bars: 5 μm.