Procedural rigging of a realistic human face

Authors

  • Nenad Šunjka Autor
  • Lidija Krstanović Fakultet tehničkih nauka, Novi Sad Autor

DOI:

https://doi.org/10.24867/06SA01Sunjka

Keywords:

Rigging, facial rigging, character rigging, artificial intelligence, computer graphics

Abstract

This paper describes the concept of procedural rigging of a realistic human face that is completely automated. The implementation of such a system, which was developed within the Maya software using the Python programming language, was also analyzed. The result is an application called Rigen which offers the user to build a rig for arbitrary character in six steps, and thus prepare it for computer animation. The steps are: creating a directory; defining character models; creating a digital skeleton; build a rig; correction of control shape objects; and assigning weights corresponding to the influence of certain bones on vertices. These steps are seemingly straightforward, if viewed from the angle of manual crew rig. Nevertheless, as such, they build a stable framework that allows the procedurality of the entire process. Complete automation of this process is reflected in the step of creating a digital skeleton using artificial intelligence. Unlike all commercial software, where it is necessary to manually create a digital skeleton for face, in this case, this is achieved automatically by using the pre-trained neural network.

References

[1] V. Kazemi, J. Sulivan, One Millisecond Face Alignment with an Ensemble of Regression Trees, Stockholm, 2015.
[2] L. Dutreve, A. Meyer, V. Orvalho, S. Bouakaz, Easy Rigging of Face by Automatic Registration and Transfer of Skinning Parameters, HAL, 2017.
[3] P. Viola, M. Jones, Robust Real-Time Face Detection, Cambridge, 2003.
[4] Besplatan model karaktera preuzet sa adrese: www.turbosquid.com, pristupljeno: 3.7.2018.

Published

2020-01-30