Compressing the Input for CNNs with the First-Order Scattering Transform

Daniele Calandriello 1 Alessandro Lazaric 1 Michal Valko 1
1 SEQUEL - Sequential Learning
Inria Lille - Nord Europe, CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille (CRIStAL) - UMR 9189
Abstract : We study the first-order scattering transform as a candidate for reducing the signal processed by a convolutional neural network (CNN). We study this transformation and show theoretical and empirical evidence that in the case of natural images and sufficiently small translation invariance, this transform preserves most of the signal information needed for classification while substantially reducing the spatial resolution and total signal size. We show that cascading a CNN with this representation performs on par with ImageNet classification models commonly used in downstream tasks such as the ResNet-50. We subsequently apply our trained hybrid ImageNet model as a base model on a detection system, which has typically larger image inputs. On Pascal VOC and COCO detection tasks we deliver substantial improvements in the inference speed and training memory consumption compared to models trained directly on the input image.
Type de document :
Communication dans un congrès
European Conference on Computer Vision, 2018, Munich, Germany
Liste complète des métadonnées

Littérature citée [13 références]  Voir  Masquer  Télécharger
Contributeur : Michal Valko <>
Soumis le : vendredi 3 mars 2017 - 18:35:50
Dernière modification le : vendredi 22 mars 2019 - 01:34:57
Document(s) archivé(s) le : mardi 6 juin 2017 - 13:15:28


Fichiers produits par l'(les) auteur(s)


  • HAL Id : hal-01482760, version 1


Daniele Calandriello, Alessandro Lazaric, Michal Valko. Compressing the Input for CNNs with the First-Order Scattering Transform. European Conference on Computer Vision, 2018, Munich, Germany. 〈hal-01482760〉



Consultations de la notice


Téléchargements de fichiers