Patent-family-89158456

The present disclosure relates to a computer-implemented method of aligning digital three-dimensional (3D) representations of an upper and lower jaw, the method comprising the steps of: generating a first digital 3D representation of the upper jaw and a second digital 3D representation of the lower jaw; inputting the first and second digital 3D representations to a first neural network configured to output a transformation for each 3D representation; applying the transformation to each of the first and second 3D representations, whereby transformed 3D representations are generated; inputting the transformed 3D representations to a second neural network configured to align the transformed 3D representations; and fine tuning the alignment such that the first and second 3D representations of the jaws are arranged in occlusion.