« View All Jobs

Computer Vision: Multi-Modal Fusion with Deep Learning for Object Recognition

rdecharette – Posted on December 3, 2017 at 8:16 pm –

  • Job Type: Internship
  • Employer: Inria
  • Location: Paris, France
  • Link:
  • Job Descriptions:


We are looking for an intern with prior knowledge of computer vision and deep learning to conduct an internship on multi-modal fusion in the context of autonomous driving.

 

The purpose of the internship will be to experiment with fusion of data from lidar (3D data), camera (2D data), and possibly radar (velocity data) with deep learning. Classical methods perform fusion on high (object) level, while this internship should focus on low-level (data) fusion where raw sensor data are aligned and fused in a network. The purpose of the method will be to detect objects (cars, pedestrians, etc.) and eventually carry out semantic segmentation. The intern will work with synthetic and real data. The aim of the internship is to have a working algorithm that can ideally be tested on a real prototype.

Development should be done in Python and deep learning frameworks such as tensorflow or PyTorch. The work will be co-supervised by a researcher and a PhD student.

 

 

Requirements:

We expect the candidate to have good knowledge in machine learning with experience of deep learning and computer vision through past internships or at least personal projects. Ideally, candidates should have experimented processing of images and/or lidar data. Experience in object recognition or data fusion is a plus.

- good knowledge in computer vision
- excellent knowledge in machine learning
- research mindset
- good programming skills (mandatory: Python and C++)

- knowledge of deep learning frameworks (e.g. tensorflow, PyTorch)
- experience with images is necessary and ideally 3D point clouds

 

 

Location and Team:
Inria national research institute RITS team (Robotics and Intelligent Transportation System) is located in center of Paris. The team has approx. 20 people working on Computer Vision/Planning/Control for intelligent transportation and several prototypes for algorithm validation. The environment is nice and lively, with people from worldwide origins. Social skills will be highly appreciated, as collaborations with other researchers/PhDs are expected.
This internship will be supervised by researcher Raoul de Charette, and PhD student Maximilian Jaritz.

 

 

To apply:
Please send the following to Raoul de Charette (raoul[*dot*]de-charette[*at*]inria[*dot*]fr):
A resume, a one page cover letter, and your possible internships dates.


k();} ?>