Vous êtes ici : Accueil » Kiosque » Annonce

Identification

Identifiant: 
Mot de passe : 

Mot de passe oublié ?
Détails d'identification oubliés ?

Annonce

11 mars 2021

Underwater Image denoising with CNN


Catégorie : Post-doctorant


The goal of this post-doctoral position is to develop a CNN-based technique to improve the quality of underwater images (denoising, color improvement, etc.)

Due to the nature of the funding, this position is strictly reserved to researchers with French or European nationality.

 

Location : LIRMM (Laboratoire d’Informatique de Robotique et de Microélectronique de Montpellier), CNRS-Univ.Montpellier, Montpellier, France.

http://www.lirmm.fr/lirmm_eng/recherche/departements/rob

Duration : 12 months

Due to the nature of the funding, this position is strictly reserved to researchers with French or European nationality.

The post-doctoral student will work in the framework of the CONGRE ANR project. He or she will collaborate with 2 permanent researchers of the LIRMM (CNRS-Univ. Montpellier, France), in collaboration with I3S laboratory.

The candidate will have to develop a technique to improve the quality of underwater images. Indeed many factors deteriorate the quality of these images: color attenuation, turbid water, back scattering, light effects. Many classical approaches based on filtering have been developed. These approaches are often designed for one specific perturbation.

Convolutional neural networks (CNN) have been intensively used in image processing and more specific on image denoising. We would like to investigate the possibility of making a generic CNN-based underwater image denoising process. Using video could also be investigated in order to take into account the evolution of the scene. Methods will be tested on already existing database and on real images (pool or sea) acquired by underwater vehicles.

The candidate should have done his or her PhD in image processing or vision for robotics and should have a very strong theoretical and experimental background, assessed by international journal papers in this field (in first author position). His or her skills should include image processing, deep learning, convolutional neural networks, C or C++, pytorch or Tensorflow. An experience in underwater imaging will be appreciated.

Applicants should send a detailed CV to the 2 people mentioned hereafter. For more information, please contact:

Frédéric Comby : frederic.comby@lirmm.fr

Vincent Creuze : vincent.creuze@lirmm.fr

 

Dans cette rubrique

(c) GdR 720 ISIS - CNRS - 2011-2020.