Vous êtes ici : Accueil » Kiosque » Annonce


Mot de passe : 

Mot de passe oublié ?
Détails d'identification oubliés ?


5 septembre 2017

GPU acceleration and 3D rendering for multiple camera pedestrian detection

Catégorie : Stagiaire

Internship objectives

Crowd analysis is an emerging topic in computer vision, being closely related to image processing, machine learning and multiple object tracking. Advances in this field benefit to a variety of problems, which may be more conceptual, such as graph modelling for collective motion analysis, or more practical, such as improving the security of specific locations such as railway stations or stadium exits.

  1. The first objective is to design and implement a parallel GPU algorithm (using the CUDA framework) for pedestrian detection through optimization of graphical models, based on the algorithm we proposed in our group [1]. The implementation of such methodology introduces new challenges in the GPU setting, like proper thread coalescing of complex arrays of structures, and optimal use of limited shared memory under massive reuse of data.
  2. A secondary objective of the internship is the creation of a simple web-based viewer of a multiple camera system and of the ground plane, based on the output of our pose estimation algorithms. This application is intended to evolve progressively into a user-friendly interface for monitoring the detections in a camera network.

For more information about the context of the internship, visit the project webpage:



Master level or equivalent, with solid background in computer science and software design. Familiarity with GPU programming and computer vision will be appreciated, as well as previous experience with WebGL or equivalent libraries.

Starting date

As soon as possible, but no later than March 2018.




[1] N. Pellicanò, E. Aldea, and S. Le Hégarat-Mascle, “Geometry-based multiple camera head detection in dense crowds,” in Proceedings of the 28th British Machine Vision Conference (BMVC) - 5th Activity Monitoring by Multiple Distributed Sensing Workshop, 2017.


Dans cette rubrique

(c) GdR 720 ISIS - CNRS - 2011-2018.