From 2D to 3D: Construction of a 3D Parametric Model for Detection of Dental Roots Shape and Position from a Panoramic Radiograph—A Preliminary Report

Abstract

Objectives. To build a 3D parametric model to detect shape and volume of dental roots, from a panoramic radiograph (PAN) of the patient. Materials and Methods. A PAN and a cone beam computed tomography (CBCT) of a patient were acquired. For each tooth, various parameters were considered (coronal and root lengths and widths): these were measured from the CBCT and from the PAN. Measures were compared to evaluate the accuracy level of PAN measurements. By using a CAD software, parametric models of an incisor and of a molar were constructed employing B-spline curves and free-form surfaces. PAN measures of teeth 2.1 and 3.6 were assigned to the parametric models; the same two teeth were segmented from CBCT. The two models were superimposed to assess the accuracy of the parametric model. Results. PAN measures resulted to be accurate and comparable with all other measurements. From model superimposition the maximum error resulted was 1.1 mm on the incisor crown and 2 mm on the molar furcation. Conclusion. This study shows that it is possible to build a 3D parametric model starting from 2D information with a clinically valid accuracy level. This can ultimately lead to a crown-root movement simulation.

Converting 2D magnetic resource imagining brain tumors to 3D structure using depth map machine learning techniques

“…Quang et al [11], The author provides an overview of DICOM computer vision translation, and also a three-dimensional concept of digitized and bridge imaging and how to assess and develop it using the VTK library. Mazzotta et al [12]. Present an overview of how to build a three-dimensional modeling of dental roots without using CBCT while decreasing the radiation rate to safeguard the patient's health and providing three-dimensional information with statistical and clinically acceptable accuracy.…”

X2Teeth: 3D Teeth Reconstruction from a Single Panoramic Radiograph

“…There have been several researches on the 3D reconstruction of a single tooth from its 2D scanning. For example, [10] models the volume of a tooth from Xrays by deforming the corresponding tooth atlas according to landmark aligning. [1,2] reconstruct a tooth from its crown photo by utilizing the surface reflectance model with shape priors.…”

“…This task is more challenging than the single tooth reconstruction, since not only tooth shapes but also spatial localization of teeth should be estimated from their 2D representation. Moreover, all the existing methods of tooth reconstruction [1,2,10] utilize ad-hoc image processing steps and handcrafted shape features. Currently, Convolutional Neural Networks (ConvNet) provide an accurate solution for single-view 3D reconstruction by discriminative learning, and have become the state-of-the-art for many photo-based benchmarks [6,8,14].…”

Geometrical modeling of complete dental shapes by using panoramic X-ray, digital mouth data and anatomical templates

“…The major disadvantage of these methods relies on the use of tomographic recording, which is not considered a routine approach for standard orthodontic treatments, due to radiation issues. For this reason, alternative proposals have been recently developed by using prebuilt libraries through statistical models [15] or by combining information derived from panoramic radiography with CAD templates [16,17]. However, very few attempts have been made up to now within the scientific community and the existent studies greatly rely on the knowledge of the specific radiographic device used to acquire PAN images.…”