US 12,460,923 B2
Systems and methods of 3D object reconstruction using a neural network
Gleb Alexandrovich Gusev, Mensdorf (LU); Vadim Bereznyuk, Moscow (RU); Kseniya Cherenkova, Luxembourg (LU); and Ilya Arzhannikov, Bertrange (LU)
Assigned to ARTEC EUROPE S.À R.L., Senningerberg (LU)
Filed by ARTEC EUROPE S.À R.L., Senningerberg (LU)
Filed on Feb. 8, 2023, as Appl. No. 18/166,396.
Application 18/166,396 is a continuation of application No. PCT/IB2021/000555, filed on Aug. 13, 2021.
Claims priority of provisional application 63/070,066, filed on Aug. 25, 2020.
Prior Publication US 2023/0184542 A1, Jun. 15, 2023
Int. Cl. G01B 11/25 (2006.01)
CPC G01B 11/2513 (2013.01) [G01B 11/2518 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method, comprising:
obtaining an image of an object, wherein the image includes a plurality of imaged elements of an imaged pattern, wherein:
the imaged pattern corresponds to a projection pattern that is projected onto a surface of the object via a slide; and
the projection pattern includes a plurality of projection elements;
providing the image of the object as an input to a neural network;
using the neural network to output a correspondence between respective pixels of the plurality of imaged elements and coordinates of the plurality of projection elements, wherein the coordinates describe the position of respective projection elements of the projection pattern on the slide; and
using the correspondence between the respective pixels of the plurality of imaged elements and the coordinates of the plurality of projection elements, reconstructing a shape of the surface of the object.