Volume Render Dental CBCT DICOM

@hkhkkh

You need ITK-SNAP volume renders, not a segmentation job - so here is exactly what I deliver and the settings I use. 1. I open each CBCT series in ITK-SNAP with the DICOM reader, not a NIfTI conversion, so the original geometry, voxel spacing and orientation survive and you can re-open the project file and see the identical scene. 2. Windowing: I set the transfer function per series instead of taking the default. For dental CBCT the useful window is narrow around the enamel/dentin step. I send you two candidate windows as PNG pairs before committing to the final render, so you pick, not me. 3. Colour maps: greyscale for the anatomical readout plus one colour-mapped version (hot / blue) for the presentation copy. 4. Orientation: I add L/R/A/P/S/I labels and keep patient orientation consistent across every render. A volume render with flipped L/R is worse than no render at all. 5. Restorations and implants create streak bands. I do not segment, but I do pick the opacity curve that keeps those bands from hiding the roots, and if a series is unusable I tell you which one and why instead of shipping a blurred render. Deliverables exactly as you listed them: - .itksnap project file with all view parameters saved - a short screen recording of the exact steps and settings (2-4 minutes, no narration needed) - high resolution PNG/TIFF exports at 4x the viewport One detail that is easy to miss and decides the whole result: CBCT series frequently arrive with an axial order ITK-SNAP reads as reversed, which puts the mandible above the maxilla in 3D. I check the slice order and the first/last slice DICOM metadata before rendering and correct the axis if it is flipped. This is the difference between a render that matches the surgeon's mental model and one that quietly misleads everyone. Send one study first. I return the complete deliverable set for it within 24 hours so you approve the look before I process the remaining studies, and I keep the same window and colour settings across the whole set so the renders are comparable.