Sparse and topological coding for visual localization of autonomous vehicles - ETIS, équipe de recherche CELL Access content directly
Conference Papers Year : 2022

Sparse and topological coding for visual localization of autonomous vehicles

Abstract

Efficient encoding of visual information is essential to the success of vision-based navigation tasks in large-scale environments. To do so, we propose in this article the Sparse Max-Pi neural network (SMP), a novel compute-efficient model of visual localization based on sparse and topological encoding of visual information. Inspired by the spatial cognition of mammals, the model uses a "topologic sparse dictionary" to efficiently compress the visual information of a landmark, allowing rich visual information to be represented with very small codes. This descriptor, inspired by the neurons in the primary visual cortex (V1), are learned using sparse coding, homeostasis and self-organising map mechanisms. Evaluated in cross-validation on the Oxford-car dataset, our experimental results show that the SMP model is competitive with the state of the art. It thus provides comparable or better performance than CoHog and NetVlad, two state-of-the-art VPR models.
Fichier principal
Vignette du fichier
2022_SAAB_v2.pdf (1.55 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03773382 , version 1 (09-09-2022)

Identifiers

  • HAL Id : hal-03773382 , version 1

Cite

Sylvain Colomer, Nicolas Cuperlier, Guillaume Bresson, Steve Pechberti, Olivier Romain. Sparse and topological coding for visual localization of autonomous vehicles. FROM ANIMALS TO ANIMATS 16: The 16th International Conference on the Simulation of Adaptive Behavior (SAB2022), Sep 2022, Cergy Pontoise, France. ⟨hal-03773382⟩
73 View
44 Download

Share

Gmail Facebook X LinkedIn More