Facial point cloud data

[Turing2022]

Dear Author, could you please explain how you converted the CBCT data into a facial point cloud dataset? If you could provide specific methods, it would be greatly appreciated. Thank you very much.

[lancesye]

CBCT Data Preprocessing: This study first extracts a three-dimensional facial model from CBCT images. Initially, threshold segmentation is applied to obtain a binary mask containing the soft tissues, and hole-filling is performed on each slice to generate a closed three-dimensional voxel region. Subsequently, the Marching Cubes algorithm is used to reconstruct a three-dimensional surface model only at the interfaces between soft tissue and air, thereby avoiding interference from boundary artifacts. The extracted surface model is then subjected to largest connected component analysis to remove isolated points, and the anterior facial region is retained through cropping. Finally, a Sinc window‑based filtering method is applied to smooth the model, improving mesh quality while preserving key facial structural features. The preprocessing results are shown in the figure.

Fig.1. Facial model extraction from CBCT. From left to right: CBCT sagittal slice, threshold segmentation with hole filling, surface reconstructed from the mask, and automatically cropped and filtered facial model.