Position: Research engineer – ShapeUp Keiki!

Position:  Research engineer at INRIA – Grenoble, MORPHEO team

Date: The position is open and the candidate can start as soon as possible. Funding is for 12 months (renewable). A start in September 2022 could be possible.

Advisors: The retained candidate will be advised by Sergi Pujades (Morpheo INRIA Grenoble, France)  and Nikolas Hesse (Swiss Children’s Rehab, University Children’s hospital Zurich, Switzerland).

Context:

The ShapeUp Keiki  project is the third project of the ongoing ShapeUp Studies series, where humans in the ages ranging from newborn to 5 year old will be studied.

The purpose of these studies is to explore and develop ways to measure health and body composition from 2D and 3D images, and optical scans. These technologies aim to take a look inside areas of the human body hidden by our skin. The study will test if new imaging machines can provide useful and detailed information about various health and wellness risks. A cohort of voluntary participants will provide the data, creating the largest and most powerful description of optical body shape and its association to body composition, metabolic markers, function, and dietary intake.

Objectives

In this project the research engineer will work on the registration of the articulated body model SMIL [1, 2] to scan data (3D point clouds) of infants. An existing code base exists, which is suitable for adults, but infants present several additional challenges. Their poses can not be directed and automatic methods of pose detection will need to be studied. The shape of the participants of the ShapeUp Studies is also different from the shape of the population used to learn the SMIL model. The goal will be to create a more generic body model that better captures the variations in shape.

 

Left: input RGB image. Right: obtained SMIL model

[1] Learning an Infant Body Model from RGB-D Data for Accurate Full Body Motion Analysis
Hesse, N., Pujades, S., Romero, J., Black, M. J., Bodensteiner, C., Arens, M., Hofmann, U. G., Tacke, U., Hadders-Algra, M., Weinberger, R., Müller-Felber, W., Schroeder, A. S.
In International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), September 2018

[2] Learning and Tracking the 3D Body Shape of Freely Moving Infants from RGB-D sequences
Hesse, N., Pujades, S., Black, M. J., Arens, M., Hofmann, U. G., Schroeder, A. S.
Transactions on Pattern Analysis and Machine Intelligence, 42 (10), Special Issue on RGB-D Vision, pp. 2540-2551, 2019.

 

Data corpus:

The  data is acquired by consortium partners. No data acquisition will be performed by the candidate in INRIA.

 

Candidate Profile:

  • A master in Computer Science or Applied Mathematics (mandatory).
  • 
Strong mathematical background – geometry  – optimization techniques
  • 
Strong coding skills (pytorch, pytorch3D)
  • 
Good Oral and written English
  • Preliminary experience in the following areas is a plus: computer vision – 3D point clouds – registration techniques – geometry processing – 3D pose estimation.

A specific section in the application letter must explain the personal experience in these areas.

Important:

– Due to the collaboration with the Hawaiian Cancer Centers, weekly group meetings are held in the evening in Europe (20h – 22h). The candidate should arrange once a week to be available at this time.

– An annual consortium meeting is held in Hawaii (usually in summer). The candidate should be available for a business travel of approx. a week.

 

How to apply:

Please send your application including

  • Mandatory: Complete CV
  • Mandatory: Letter of motivation (at most one page) – briefly describing the personal experience in the relevant areas (see Candidate Profile).
  • Mandatory: Degrees and lists of grades (translated to English or French)
  • Mandatory: Name and e-mail address of two references
  • Topic of Master thesis & Thesis and reports if available

at the jobin portal: https://jobs.inria.fr/public/classic/en/offres/2022-04684

NOTE: only complete applications on jobin will be considered.

Comments are closed.