Bayesian Classification of Affordances from RGB Images
DOI:
https://doi.org/10.26754/jjii3a.20227002Abstract
We propose a Bayesian classification model using Deep Ensembles (DE) and MC-Dropout (MC-D) to predict affordances from RGB images. Our Bayesian model obtains a higher performance than previous works and captures the aleatoric and epistemic uncertainty, showing consistency with the type of objects and scenarios.
Display downloads
Downloads
Published
2022-07-18
Issue
Section
Artículos (Tecnologías de la Información y las Comunicaciones)
How to Cite
Mur, L., & Martínez-Cantín, R. (2022). Bayesian Classification of Affordances from RGB Images. Jornada De Jóvenes Investigadores Del I3A, 10. https://doi.org/10.26754/jjii3a.20227002