PaSCo: Urban 3D Panoptic Scene Completion with Uncertainty Awareness - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Preprints, Working Papers, ... Year : 2023

PaSCo: Urban 3D Panoptic Scene Completion with Uncertainty Awareness

Abstract

We propose the task of Panoptic Scene Completion (PSC) which extends the recently popular Semantic Scene Completion (SSC) task with instance-level information to produce a richer understanding of the 3D scene. Our PSC proposal utilizes a hybrid mask-based technique on the nonempty voxels from sparse multi-scale completions. Whereas the SSC literature overlooks uncertainty which is critical for robotics applications, we instead propose an efficient ensembling to estimate both voxel-wise and instance-wise uncertainties along PSC. This is achieved by building on a multi-input multi-output (MIMO) strategy, while improving performance and yielding better uncertainty for little additional compute. Additionally, we introduce a technique to aggregate permutation-invariant mask predictions. Our experiments demonstrate that our method surpasses all baselines in both Panoptic Scene Completion and uncertainty estimation on three large-scale autonomous driving datasets. Our code and data are available at https://astra-vision.github.io/PaSCo .
Fichier principal
Vignette du fichier
2312.02158.pdf (16.16 Mo) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04324930 , version 1 (05-12-2023)

Identifiers

  • HAL Id : hal-04324930 , version 1

Cite

Anh-Quan Cao, Angela Dai, Raoul de Charette. PaSCo: Urban 3D Panoptic Scene Completion with Uncertainty Awareness. 2023. ⟨hal-04324930⟩
104 View
35 Download

Share

Gmail Mastodon Facebook X LinkedIn More