Facial Landmarks Localization Estimation by Cascaded Boosted Regression

Abstract : Accurate detection of facial landmarks is very important for many applications like face recognition or analysis. In this paper we describe an efficient detector of facial landmarks based on a cascade of boosted regressors of arbitrary number of levels. We define as many regressors as landmarks and we train them separately. We describe how the training is conducted for the series of regressors by supplying training samples centered on the predictions of the previous levels. We employ gradient boosted regression and evaluate three different kinds of weak elementary regressors, each one based on Haar features: non parametric regressors, simple linear regressors and gradient boosted trees. We discuss trade-offs between the number of levels and the number of weak regressors for optimal detection speed. Experiments performed on three datasets suggest that our approach is competitive compared to state-of-the art systems regarding precision, speed as well as stability of the prediction on video streams.
Type de document :
Communication dans un congrès
International Conference on Computer Vision Theory and Applications (VISAPP 2013), Feb 2013, Barcelona, Spain. 2013
Liste complète des métadonnées

Littérature citée [18 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-00818808
Contributeur : Alexey Ozerov <>
Soumis le : lundi 29 avril 2013 - 11:33:10
Dernière modification le : vendredi 16 février 2018 - 11:40:02
Document(s) archivé(s) le : mardi 4 avril 2017 - 01:35:20

Fichier

article-Visapp.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-00818808, version 1

Collections

UGA

Citation

Louis Chevallier, Jean-Ronan Vigouroux, Alix Goguey, Alexey Ozerov. Facial Landmarks Localization Estimation by Cascaded Boosted Regression. International Conference on Computer Vision Theory and Applications (VISAPP 2013), Feb 2013, Barcelona, Spain. 2013. 〈hal-00818808〉

Partager

Métriques

Consultations de la notice

465

Téléchargements de fichiers

1027