On regression losses for deep depth estimation - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2018

On regression losses for deep depth estimation

À propos des fonctions de pénalité pour l'estimation de profondeur par réseaux profonds

Résumé

Depth estimation from a single monocular image has reached great performances thanks to recent works based on deep networks. However, as various choices of losses, architectures and experimental conditions are proposed in the literature, it is difficult to establish their respective influence on the performances. In this paper we propose an in-depth study of various losses and experimental conditions for depth regression, on NYUv2 dataset. From this study we propose a new network for depth estimation combining an encoder-decoder architecture with an adversarial loss. This network reaches top ones state of the art on NUYv2 dataset while being simpler to train in a single phase.
Fichier principal
Vignette du fichier
DTIS18200.1541512021_preprint.pdf (2.46 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01925321 , version 1 (16-11-2018)

Identifiants

Citer

Marcela Carvalho, Bertrand Le Saux, Pauline Trouvé-Peloux, Andrés Almansa, Fréderic Champagnat. On regression losses for deep depth estimation. ICIP 2018, Oct 2018, ATHENES, Greece. ⟨10.1109/ICIP.2018.8451312⟩. ⟨hal-01925321⟩
122 Consultations
1666 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More