PhD position offer: Dense Volumetric 4D Modelling of Moving Shapes

Context
The Morpheo team is working on the capture and modelling of moving shapes using visual cues. Recent activities in the team are considering the combination of video cameras with X-ray imaging in order to enable combined surface and inside shape modelling. These activities build on the Kinovis platform that enables the simultaneous acquisition of colour and X-ray image sequences. The PhD proposal is part of this effort and funded by a research project that involves in particular the shoe industry. The main objective of the project is to design and experiment new dense 4D models of the feet that allow for the combined analysis of internal and external feet and shoe dynamics.

PhD Objectives
The overall objective of this PhD project is to develop new dense models for human motion analysis, especially articulated motions as with feet and hands. Using combined X-ray and colour cameras, it is possible to capture both surface motion (skin/shoe) and information on the skeletal structure, as illustrated below (left). A dense volumetric model can then be reconstructed over time (right) using both sources of information.

  
Left: combined capture with multiple video and X-ray. Right: reconstructed dense 3D model over time.

Little research has yet been carried out about the complex dynamics between the skin and the underlying skeleton. Soft and elastic tissues such as muscles, fat, and skin are indeed usually ignored or considered as an adjustment variable. The main goal of this PhD is to study how the soft tissues (skin, muscles, fat) and the skeleton interact and to propose models enabling inference from interactions in both directions. Machine learning approaches will be considered to this end.

Candidate profile
The PhD candidate should hold a master’s degree in computer science. Very good background in computer vision, 3D modelling, or medical imaging are expected. Knowledge in biomechanics or sports science is a plus. The candidate will be co-supervised by Julien Pansiot and Edmond Boyer at Inria Grenoble, France.

Inria Grenoble
Inria is a leading French research centre in computer science, with an international culture – the English language being widely adopted. The Grenoble centre is located at the heart of the French Alps, a very dynamic region for new technologies offering a large range of recreational activities.

Informal inquires can be addressed to julien.pansiot|edmond.boyer@inria.fr. Please upload your application, quoting the project CaMoPi, on the team website: http://morpheo.inrialpes.fr/job-applications.

References

  1. Julien Pansiot and Edmond Boyer. 3D Imaging from Video and Planar Radiography. In International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), Athens, October 2016.
  2. A. Baudoinet al., Parametric subject-specific model for in vivo 3D reconstruction using bi-planar X- rays: application to the upper femoral extremity. Medical & Biological Engineering & Computing 46, 799–805, 2008.