ADAPT Project – Innovation award at SOFMER 2019

The Interreg ADAPT project ( aims to define innovative mobility solutions for people with disabilities. A power wheelchair simulator has been co-created with the medical staff of the Pôle Saint Hélier, a rehabilitation center in Rennes. Our joint work was rewarded by LNA Santé during the SOFMER Congress 2019.

Lire la suite

Samuel Felton

Désolé, cet article est seulement disponible en anglais américain. Pour le confort de l’utilisateur, le contenu est affiché ci-dessous dans une autre langue. Vous pouvez cliquer le lien pour changer de langue active.

body #nav-below{ display: none;}

Samuel Felton

Phd student, Université Rennes 1

Email : samuel.felton @
Address :
IRISA, INRIA Rennes – Bretagne Atlantique
Campus Universitaire de Beaulieu
35042 Rennes, France
Assistant : 02 99 84 22 52 (Hélène de La Ruée)




I received a double master’s degree (computer science engineering – research-oriented computer science) from the INSA Rennes engineering school in 2019. My last master’s year has been heavily focused on the areas of computer vision and machine learning. Beforehand, I studied at the IUT du Limousin and received a « DUT informatique », a technical computer science degree.

I have completed two internships related to machine learning and vision problems.

  • The first one was done at Robovision in 2018 and aimed at improving the performance of a neural network trained to regress the orientation of tulip bulbs that were to be prepared for plantation by a robotic arm.  I worked with both point clouds and raw images of the bulbs to improve orientation regression.
  • The second one took place at Inria in 2019 and concluded my master’s degree.  The subject of this internship was to apply deep learning to the problem of Visual servoing. It is the foundation on which my PhD started.

The topic of my PhD is the following: « Deep learning for visual servoing » and is under the supervison of Elisa Fromont (LACODAM team) and Eric Marchand (Rainbow team). I will be exploring how to apply deep learning techniques to perform efficient visual servoing. As of now, I am focusing on performing visual servoing in an end-to-end fashion and regressing the camera velocity directly instead of estimating the camera poses (that can be used in Position-Based Visual Servoing).