Using a Waffle Iron for Automotive Point Cloud Semantic Segmentation - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2024

Using a Waffle Iron for Automotive Point Cloud Semantic Segmentation

Résumé

Semantic segmentation of point clouds in autonomous driving datasets requires techniques that can process large numbers of points efficiently. Sparse 3D convolutions have become the de-facto tools to construct deep neural networks for this task: they exploit point cloud sparsity to reduce the memory and computational loads and are at the core of today's best methods. In this paper, we propose an alternative method that reaches the level of state-of-the-art methods without requiring sparse convolutions. We actually show that such level of performance is achievable by relying on tools a priori unfit for large scale and high-performing 3D perception. In particular, we propose a novel 3D backbone, WaffleIron, made almost exclusively of MLPs and dense 2D convolutions and present how to train it to reach high performance on SemanticKITTI and nuScenes. We believe that WaffleIron is a compelling alternative to backbones using sparse 3D convolutions, especially in frameworks and on hardware where those convolutions are not readily available. The code is available at https://github.com/valeoai/WaffleIron.
Fichier principal
Vignette du fichier
Puy_Using_a_Waffle_Iron_for_Automotive_Point_Cloud_Semantic_Segmentation_ICCV_2023_paper.pdf (608.72 Ko) Télécharger le fichier
Origine Fichiers éditeurs autorisés sur une archive ouverte

Dates et versions

hal-04496531 , version 1 (08-03-2024)

Identifiants

Citer

Gilles Puy, Alexandre Boulch, Renaud Marlet. Using a Waffle Iron for Automotive Point Cloud Semantic Segmentation. 2023 IEEE/CVF International Conference on Computer Vision (ICCV 2023), Oct 2023, Paris, France. pp.3356-3366, ⟨10.1109/ICCV51070.2023.00313⟩. ⟨hal-04496531⟩
17 Consultations
5 Téléchargements

Altmetric

Partager

Gmail Mastodon Facebook X LinkedIn More