DIR (Dental Image repository) present a 3D dental Database for benchmarking detection, segmentation and classification. Each point in the 3D point cloud has been segmented and classified, allowing point evaluation.
This dataset contains 3D data from dental plaster models. Approximatively 60.000 points are acquire per piece. the acquisition was made on image laboratory of National University of Colombia. The dataset have five PLY models wich each file contain a list of (X, Y, Z, label)points, where XYZ correspond to geometric object and label correspond to the object identifier obtained.
Eight views were captured in order to construct each digital model, with the scanner Konica Minolta Vivid 9i. Tilted at 45 degrees, achieving good detail within the object of interest.
Annotation has been carried out in a manually way. Then, a label and class is assigned to each segmented object acoording to ISO System by the World Health Organization.
This database is avaliable under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4)
Insert the following citation in any scientific or technical publication whenever this dataset was used to get the results:
J.D. Tamayo-Quintero, J.B. Gómez-Mendoza "Digital Dental Three Dimensional Database: a 3D dataset for benchmarking digital dentistry, manual annotations and ground truth".