Learning Non-linear SVM in Input Space for Image Classification - Archive ouverte HAL Accéder directement au contenu
Rapport Année : 2014

Learning Non-linear SVM in Input Space for Image Classification

Gaurav Sharma
  • Fonction : Auteur
  • PersonId : 940543
Patrick Pérez
  • Fonction : Auteur
  • PersonId : 1022281

Résumé

The kernel trick enables learning of non-linear decision functions without having to explicitly map the original data to a high dimensional space. However, at test time, it requires evaluating the kernel with each one of the support vectors, which is time consuming. We propose a novel approach for learning non-linear support vector machine (SVM) corresponding to commonly used kernels in computer vision, namely (i) Histogram Intersection, (ii) χ2, (ii) Radial Basis Function (RBF) and (iv) RBF with χ2 distance, without using the kernel trick. The proposed classifier incorporates non-linearity while maintaining O(D) testing complexity (for D- dimensional space), compared to O(D × Nsv) (for Nsv number of support vectors) when using the kernel trick. We also promote the idea that such efficient non-linear classifier, combined with simple image encodings, is a promising direction for image classification. We validate the proposed method with experiments on four challenging image classification datasets. It achieves similar performance w.r.t. kernel SVM and recent explicit feature mapping method while being significantly faster and memory efficient. It obtains competitive performance while being an order of magnitude faster than the state-of-the-art Fisher Vector method and, when combined with it, consistently improves performance with a very small additional computation cost.
Fichier principal
Vignette du fichier
nsvm_pami.pdf (1.47 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-00977304 , version 1 (10-04-2014)
hal-00977304 , version 2 (10-12-2014)

Identifiants

  • HAL Id : hal-00977304 , version 1

Citer

Gaurav Sharma, Frédéric Jurie, Patrick Pérez. Learning Non-linear SVM in Input Space for Image Classification. 2014. ⟨hal-00977304v1⟩

Collections

UGA
561 Consultations
848 Téléchargements

Partager

Gmail Facebook X LinkedIn More