Discrimination of visual pedestrians data by combining projection and prediction learning - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2014

Discrimination of visual pedestrians data by combining projection and prediction learning

Abstract

PROPRE is a generic and semi-supervised neural learning paradigm that extracts meaningful concepts of multimodal data flows based on predictability across modalities. It consists on the combination of two computational paradigms. First, a topological projection of each data flow on a self-organizing map (SOM) to reduce input dimension. Second, each SOM activity is used to predict activities in all other SOMs. Predictability measure, that compares predicted and real activities, is used to modulate the SOM learning to favor mutually predictable stimuli. In this article, we study PROPRE applied to a classical visual pedestrian data classification task. The SOM learning modulation introduced in PROPRE improves significantly classification performance.
Fichier principal
Vignette du fichier
article.pdf (270.49 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01061654 , version 1 (08-09-2014)

Identifiers

  • HAL Id : hal-01061654 , version 1

Cite

Mathieu Lefort, Alexander Gepperth. Discrimination of visual pedestrians data by combining projection and prediction learning. European Symposium on artificial neural networks (ESANN), Apr 2014, Bruges, Belgium. ⟨hal-01061654⟩
136 View
45 Download

Share

Gmail Facebook X LinkedIn More