Scene context is more than a Bayesian prior: Competitive vehicle detection with restricted detectors

Abstract : We present a new approach for making use of scene or situation context in object detection, aiming for state-of-the-art performance while dramatically reducing computa-tional cost. While existing approaches are inspired by Bayes' rule, training context-independent detectors and combining them with context priors in hindsight, we propose to integrate these context priors already into detector design, through algorithmic choices and/or pre-selection of training examples. Although such restricted detectors will, as a consequence, be valid only in regions compatible with context priors, the corresponding simplification of the object-vs-background de-cision problem will lead to reduced computation time and/or increased detection performance. We verify this experimentally by analyzing vehicle detection performance in a realistically simulated inner-city environment where context priors are defined by a road surface mask obtained from the simulation tool. Comparing a restricted detector, based on horizontal gradient detection refined by neural network confirmation, to a generic HOG+SVM-based approach taking into account the road context prior, we show that the restricted detector shows superior vehicle detection performance at a vastly reduced computational cost. We show qualitative results that permit the conclusion that the restricted detector will perform well on real-world scenes if appropriate road context priors are available.
Type de document :
Communication dans un congrès
IEEE International Symposium on Intelligent Vehicles(IV), May 2014, Detroit, United States. pp.1358 - 1364, 2014, 〈10.1109/IVS.2014.6856542〉
Liste complète des métadonnées

Littérature citée [19 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01098707
Contributeur : Alexander Gepperth <>
Soumis le : dimanche 28 décembre 2014 - 16:05:13
Dernière modification le : jeudi 11 janvier 2018 - 06:19:24
Document(s) archivé(s) le : dimanche 29 mars 2015 - 10:15:42

Fichier

hecht2014scene.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Collections

Citation

Thomas Hecht, Mrinal Mohit, Egor Sattarov, Alexander Gepperth. Scene context is more than a Bayesian prior: Competitive vehicle detection with restricted detectors. IEEE International Symposium on Intelligent Vehicles(IV), May 2014, Detroit, United States. pp.1358 - 1364, 2014, 〈10.1109/IVS.2014.6856542〉. 〈hal-01098707〉

Partager

Métriques

Consultations de la notice

1111

Téléchargements de fichiers

2838