Vous êtes ici : Accueil » Kiosque » Annonce

Identification

Identifiant: 
Mot de passe : 

Mot de passe oublié ?
Détails d'identification oubliés ?

Annonce

9 janvier 2018

Face Recognition Using Reduced Models


Catégorie : Stagiaire


Classification based on deep learning has proven its ability to learn adequate metric [1] and scale towards complex problem to reach human accuracy in tasks as subtle as face recognition [2]. This internship will follow on an ongoing effort in our lab to understand how to build a reduced model able to build meaningful representation able to cope with poor registration.

A successful candidate will be (medium to advanced) knowledgeable in deep learning and image processing and skilled in coding with Python using keras or tensorflow.

An opening for a PhD program is possible.

How to apply: send a CV, motivation letter and followed courses with notes to : eva.dokladalova@esiee.fr

Hosting Lab: LIGM, Laboratoire d'Informatique Gaspard-Monge UMR 8049, internship will be located at ESIEE Paris, Computer Science departement

Collaboration: Center of Mathematical Morphology, Mines-ParisTech, Fontainebleau

Internships advisors: Petr Dokladal, PhD., HDR, Eva Dokladalova, PhD.

Tuition: ~550€/month

Duration: 4-6 months

 

Recently, classification based on deep learning has proven its ability to learn adequate metric [1] and scale towards complex problem to reach human accuracy in tasks as subtle as face recognition [2]. Whereas this accuracy has been reached with automatically learnt features the resulting model is huge and the processing chain complex. The complexity of the processing chain is induced by the necessity to compensate for imperfect registration (translation, slant or rotation) prior to object recognition.

This internship will follow on an ongoing effort in our lab to understand how to build a reduced model able to build meaningful representation able to cope with poor registration.

The internship will start by a literature survey, and experimenting on simple geometrical shapes, progressively going towards building more complex models applied finally to recognition of faces.

A successful candidate will be (medium to advanced) knowledgeable in deep learning and image processing and skilled in coding with Python using keras or tensorflow.

An opening for a PhD program is possible.

How to apply:

Send a CV, motivation letter and followed courses with notes to : eva.dokladalova@esiee.fr

Hosting Lab:

LIGM, Laboratoire d'Informatique Gaspard-Monge UMR 8049, internship located at ESIEE Paris, Computer Science departement

Collaboration:

Center of Mathematical Morphology, Mines-ParisTech, Fontainebleau

Internships advisors:

Petr Dokladal, PhD., HDR, Eva Dokladalova, PhD.

Tuition:

~550€/month

Duration:

4-6 months

References:

  1. A. Krizhevsky, I. Sutskever and G. E. Hinton, ImageNet Classification with Deep Convolutional Neural Networks, Advances in Neural Information Processing Systems 25, F. Pereira and C. J. C. Burges and L. Bottou and K. Q. Weinberger, pp. 1097–1105, 2012, http://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf

  2. Y. Taigman, M. Yang, MA. Ranzato and L. Wolf, DeepFace: Closing the Gap to Human-Level Performance in Face Verification, CVPR IEEE Conf, June 2014

Dans cette rubrique

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