BEVContrast: Self-Supervision in BEV Space for Automotive Lidar Point Clouds - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

BEVContrast: Self-Supervision in BEV Space for Automotive Lidar Point Clouds

Résumé

We present a surprisingly simple and efficient method for self-supervision of 3D backbone on automotive Lidar point clouds. We design a contrastive loss between features of Lidar scans captured in the same scene. Several such approaches have been proposed in the literature from PointConstrast, which uses a contrast at the level of points, to the state-of-the-art TARL, which uses a contrast at the level of segments, roughly corresponding to objects. While the former enjoys a great simplicity of implementation, it is surpassed by the latter, which however requires a costly pre-processing. In BEVContrast, we define our contrast at the level of 2D cells in the Bird's Eye View plane. Resulting cell-level representations offer a good trade-off between the point-level representations exploited in PointContrast and segment-level representations exploited in TARL: we retain the simplicity of PointContrast (cell representations are cheap to compute) while surpassing the performance of TARL in downstream semantic segmentation.
Fichier principal
Vignette du fichier
2310.17281.pdf (9.83 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04492004 , version 1 (06-03-2024)

Identifiants

Citer

Corentin Sautier, Gilles Puy, Alexandre Boulch, Renaud Marlet, Vincent Lepetit. BEVContrast: Self-Supervision in BEV Space for Automotive Lidar Point Clouds. International Conference on 3D Vision 2024, Mar 2024, Davos, Switzerland. ⟨hal-04492004⟩
9 Consultations
1 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More