Annonce

Les commentaires sont clos.

Deformable objects shape tracking using Neural Networks

20 Janvier 2022


Catégorie : Stagiaire


Title: Deformable objects shape tracking using Neural Networks

Knowing the deformable 3D shape of an object is a crucial step for different applications such as medical and robotics. In Institut Pascal, we are working on developing novel vision-based algorithms and methods for the 3D reconstruction and real-time tracking of flexible surfaces for different types of materials and deformation. For this internship, we aim to include more learning techniques, Graph Neural Networks, to track the shape of deformable object.

The goal of this internship is: generating the dataset of deformable objects for training the network using related tools (Pytorch3d, Blender, etc.); developing the general Graph Neural Networks for deformable objects shape tracking using this dataset

The intern will work within SoftManBot project that contains many students and international connections. Successful outputs of this internship will lead to a scientific publication in one of the vision or robotics conferences which is a plus for the intern for his/her future academic career.

 

Required skills:

Coding: Python, Pytorch, and Matlab,

Knowledge: linear algebra, deep learning, computer vision.

 

Supervisors: Reza Shetab, Erol Ozgur , Mohammad Alkhatib, Youcef Mezouar (firstname.lastname@sigma-clermont.fr)

Starting date of the internship: between February 1st and April 1st

Duration of the internship: 5 to 6 months

Desired Level of student: Master 2 or last year of engineering school

Host laboratory: Institut Pascal, Campus des Cézeaux, 63178 Aubière

Application deadline: February 28, 2022