Skip to Main content Skip to Navigation
New interface
Conference papers

DUNE: Deep UNcertainty Estimation for tracked visual features

Katia Katia Lillo 1 Andrea de Maio 2 Simon Lacroix 2 Amaury Nègre 1, 3 Michèle Rombaut 1 Nicolas Marchand 1 Vercier Nicolas Vercier 4 
2 LAAS-RIS - Équipe Robotique et InteractionS
LAAS - Laboratoire d'analyse et d'architecture des systèmes
3 GIPSA-Services - GIPSA-Services
GIPSA-lab - Grenoble Images Parole Signal Automatique
Abstract : Uncertainty estimation of visual feature is essential for vision-based systems, such as visual navigation. We show that errors inherent to visual tracking, in particular using KLT tracker, can be learned using a probabilistic loss function to estimate the covariance matrix on each tracked feature position. The proposed system is trained and evaluated on synthetic data, as well as on real data, highlighting good results in comparison to the state of the art. The benefits of the tracking uncertainty estimates are illustrated for visual motion estimation.
Complete list of metadata
Contributor : Nicolas Marchand Connect in order to contact the contributor
Submitted on : Saturday, November 5, 2022 - 12:32:27 PM
Last modification on : Monday, November 7, 2022 - 10:49:17 AM


Files produced by the author(s)


  • HAL Id : hal-03790281, version 1


Katia Katia Lillo, Andrea de Maio, Simon Lacroix, Amaury Nègre, Michèle Rombaut, et al.. DUNE: Deep UNcertainty Estimation for tracked visual features. IPAS 2022 - 5th IEEE International Conference on Image Processing, Applications and Systems (IPAS 2022), Dec 2022, Genova, Italy. ⟨hal-03790281⟩



Record views


Files downloads