Missions

The postdoctoral researcher will have to take charge of the literature on classical source separation (tensor decomposition, non-negative matrix factorization) and compare these approaches to an unsupervised separation using variational autoencoders. The design of experiments to compare these approaches and their publication in a conference or journal of the source separation community are also envisaged.

Activités

Bibliographic research, software development, implementation of an experimental plan, dissemination of knowledge.

Compétences

Skills/knowledge: Neural networks, numerical optimization, signal processing.
Expertise: Python development, particularly PyTorch. Version control (Git).
Software: Autonomy

Contexte de travail

Come join the CREATIS laboratory (Center for Research in Image Acquisition and Processing for Health), we are a multidisciplinary laboratory with a broad spectrum of skills in medical imaging and playing a major role in the field of health technologies. It is renowned for its expertise in magnetic resonance imaging, ultrasound, X-rays and optics, with skills in physics, mathematics, computer science and instrumentation. Thus, CREATIS has all the skills necessary to optimize each step of the imaging process from acquisition to image analysis and medical interpretation. Thanks to a long and close collaboration with hospitals, CREATIS contributes to the development of personalized and predictive medicine of the future.

Contraintes et risques

Not applicable