Method for automatic segmentation of a dental arch

1 . A method for automatic segmentation of a dental arch, comprising:

acquiring a three-dimensional surface of the dental arch, in order to obtain a three-dimensional representation of the dental arch in a three-dimensional space, said three-dimensional representation comprising a set of three-dimensional points, referred to as vertices, forming vertices of polygonal, preferably triangular, faces;

generating two-dimensional virtual views from the three-dimensional representation, comprising a step of determining the characteristics of the virtual views comprising a sub-step of determining a wireframe representing the general shape of the dental arch and a sub-step of determining the characteristics of the virtual views by selecting virtual views distributed along the wireframe and directed towards the wireframe;

projecting the three-dimensional representation on each two-dimensional virtual view, configured to obtain, for each virtual view, an image representing each vertex and each polygonal face visible on the virtual view;

processing each image by a previously trained deep learning network, associating, with each pixel of each image, a probability vector, each index of the vector representing the probability of said pixel belonging to a class of dental tissues, from among multiple classes of dental tissues of a same size as the probability vector;

carrying out inverse projection of each image so as to assign to each vertex of the three-dimensional representation one pixel for each image on which the vertex appears and to which it corresponds, and assigning to each vertex the probability vector(s) associated with said one or more pixels; and,

determining, for each vertex, the dental tissue class to which said vertex most probably belongs based on the probability vector(s) assigned to said vertex.

2 . The method for automatic segmentation of a dental arch according to claim 1 , further comprising, prior to processing each image by the learning network, assigning each pixel of each image at least one discriminatory value, said discriminatory value being representative of a characteristic of the vertex when said pixel corresponds to a vertex on the virtual view, and to an interpolation of the characteristics of the vertices of the polygonal face when said pixel corresponds to a polygonal face on the virtual view.

3 . The automatic segmentation method according to claim 2 , wherein the discriminatory value may be of a value type selected from the following list of value types:

a vertex RGB value obtained during the acquisition of a three-dimensional surface;

a value of three-dimensional curvature at the vertex;

a distance value between the vertex and an optical center of the virtual view on which it projects;

an angle between a normal of the vertex and a direction of sight of the virtual view.

4 . The method for automatic segmentation of a dental arch according to claim 1 , wherein each virtual view is defined by an optical center (Co) comprised in the three-dimensional space, and by a picture-taking direction along a picture-taking axis.

5 . The method for automatic segmentation of a dental arch according to claim 1 , wherein the number of two-dimensional virtual views generated is between 30 and 90 views, preferably between 50 and 70 views.

6 . The method for automatic segmentation of a dental arch according to claim 1 , wherein the step of determining, for each vertex, the dental tissue class to which it most probably belongs comprises the execution of a graph cut algorithm taking as a parameter for each vertex said one or more probability vectors assigned to said vertex.

7 . A computing device comprising a non-transitory computer readable storage medium having program instructions stored therein, the instructions being executable by at least one processing core of a processing unit to cause the processing unit to perform a method for automatic segmentation of a dental arch, the device comprising:

a computing module of the program instructions for acquiring a three-dimensional surface of the dental arch, configured to obtain a three-dimensional representation of the dental arch in a three-dimensional space, said three-dimensional representation comprising a set of three-dimensional points, referred to as vertices, forming vertices of polygonal, preferably triangular, faces;

a computing module of the program instructions for generating two-dimensional virtual views from the three-dimensional representation, configured to determine the characteristics of the virtual views by determining a wireframe representing the general shape of the dental arch and by determining the characteristics of the virtual views by selecting virtual views distributed along the wireframe and directed towards the wireframe;

a computing module of the program instructions for projecting the three-dimensional representation on each two-dimensional virtual view, configured to obtain, for each virtual view, an image representing each vertex and each polygonal face visible on the virtual view;

a computing module of the program instructions for processing each image by a previously trained deep learning network, associating, with each pixel of each image, a probability vector, each index of the vector representing the probability of said pixel belonging to a class of dental tissues, from among different classes of dental tissues of a same size as the probability vector;

a computing module of the program instructions for carrying out inverse projection of each image so as to assign to each vertex of the three-dimensional representation one or more pixels of the images wherein the vertex appears and to which it corresponds, and assigning to each vertex the probability vector(s) associated with said one or more pixels; and,

a computing module of the program instructions for determining, for each vertex, the dental tissue class to which said vertex most probably belongs based on the probability vector(s) assigned to said vertex.