Learning the link between surface and anatomic models of the human body

This is a joint project between the MORPHEO INRIA  team and the company ANATOSCOPE which specialises in human anatomy modeling. The main goal of this project is to link two different types of human body representations. On one hand,  external representations with visual human body models [1,2,3] as built from visual observations, and on the other hand intrinsic representations with anatomic models [4,6] as built using MRI or CT scans. The interest is to be able to infer one representation from the other, for instance to infer the anatomy inside a body given its surface observation.
The objective of this internship is to establish the link between the representations of the surface models [1,2] and the anatomic models [4]. As a starting point, the registrations of surface and anatomic models will be provided [5]. The candidate will leverage machine learning techniques, in order to characterize the relationships between i) the shape space parametrization of SMPL and the anatomic parameters, as well as ii) a free form human shape in SMPL topology and the anatomic parameters. The solutions should provide a fast, accurate and efficient method relating both models. If time allows, the relations between the pose dependent deformations and the anatomic model parameters will also be studied.

The project will be supervised by Sergi Pujades (MORPHEO), Edmond Boyer (MORPHEO) and François Faure (ANATOSCOPE).

Student profile

  • Master student – preferably in Computer Science or Applied Mathematics.
  • Creative and highly motivated
  • Solid programming skills, Python, C++ and/or Matlab
  • Solid mathematics knowledge in linear algebra, geometry, and statistics.
  • Fluent English or French spoken.
  • Prior courses or knowledge in the areas of computer vision, medical imaging, mesh processing,
  • Machine learning or Bayesian inference is a plus

Duration: 4-6 months

Start date: March  2019.

How to apply
Please apply on the INRIA site  and only there. For more information contact sergi.pujades@inria.fr.

[1] Loper, Matthew, Naureen Mahmood, Javier Romero, Gerard Pons-Moll, and Michael J. Black. “SMPL: A skinned multi-person linear model.” ACM Transactions on Graphics (TOG) 34, no. 6 (2015): 248.
[2] Anguelov, Dragomir, Praveen Srinivasan, Daphne Koller, Sebastian Thrun, Jim Rodgers, and James Davis. “SCAPE: shape completion and animation of people.” In ACM transactions on graphics (TOG), vol. 24, no. 3, pp. 408-416. ACM, 2005.
[3] Robinette, Kathleen M., Hans Daanen, and Eric Paquet. “The CAESAR project: a 3-D surface anthropometry survey.” In 3-D Digital Imaging and Modeling, 1999. Proceedings. Second International Conference on, pp. 380-386. IEEE, 1999.
[4] Bauer, Armelle, Federico Ulliana, Ali-Hamadi Dicko, Benjamin Gilles, Olivier Palombi, and François Faure. “MyCorporisFabrica: making anatomy easy.” In ACM SIGGRAPH 2014 Studio, p. 16. ACM, 2014.
[5] Ali-Hamadi, Dicko, Tiantian Liu, Benjamin Gilles, Ladislav Kavan, François Faure, Olivier Palombi, and Marie-Paule Cani. “Anatomy transfer.” ACM Transactions on Graphics (TOG) 32, no. 6 (2013): 188.
[6] Kadleček, Petr, Alexandru-Eugen Ichim, Tiantian Liu, Jaroslav Křivánek, and Ladislav Kavan. “Reconstructing personalized anatomical models for physics-based body animation.” ACM Transactions on Graphics (TOG) 35, no. 6 (2016): 213.

Comments are closed.