MinkOcc: Towards real-time label-efficient semantic occupancy prediction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sze, Samuel, De Martini, Daniele, Kunze, Lars
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908299008409600
author Sze, Samuel
De Martini, Daniele
Kunze, Lars
author_facet Sze, Samuel
De Martini, Daniele
Kunze, Lars
contents Developing 3D semantic occupancy prediction models often relies on dense 3D annotations for supervised learning, a process that is both labor and resource-intensive, underscoring the need for label-efficient or even label-free approaches. To address this, we introduce MinkOcc, a multi-modal 3D semantic occupancy prediction framework for cameras and LiDARs that proposes a two-step semi-supervised training procedure. Here, a small dataset of explicitly 3D annotations warm-starts the training process; then, the supervision is continued by simpler-to-annotate accumulated LiDAR sweeps and images -- semantically labelled through vision foundational models. MinkOcc effectively utilizes these sensor-rich supervisory cues and reduces reliance on manual labeling by 90\% while maintaining competitive accuracy. In addition, the proposed model incorporates information from LiDAR and camera data through early fusion and leverages sparse convolution networks for real-time prediction. With its efficiency in both supervision and computation, we aim to extend MinkOcc beyond curated datasets, enabling broader real-world deployment of 3D semantic occupancy prediction in autonomous driving.
format Preprint
id arxiv_https___arxiv_org_abs_2504_02270
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MinkOcc: Towards real-time label-efficient semantic occupancy prediction
Sze, Samuel
De Martini, Daniele
Kunze, Lars
Computer Vision and Pattern Recognition
Robotics
Developing 3D semantic occupancy prediction models often relies on dense 3D annotations for supervised learning, a process that is both labor and resource-intensive, underscoring the need for label-efficient or even label-free approaches. To address this, we introduce MinkOcc, a multi-modal 3D semantic occupancy prediction framework for cameras and LiDARs that proposes a two-step semi-supervised training procedure. Here, a small dataset of explicitly 3D annotations warm-starts the training process; then, the supervision is continued by simpler-to-annotate accumulated LiDAR sweeps and images -- semantically labelled through vision foundational models. MinkOcc effectively utilizes these sensor-rich supervisory cues and reduces reliance on manual labeling by 90\% while maintaining competitive accuracy. In addition, the proposed model incorporates information from LiDAR and camera data through early fusion and leverages sparse convolution networks for real-time prediction. With its efficiency in both supervision and computation, we aim to extend MinkOcc beyond curated datasets, enabling broader real-world deployment of 3D semantic occupancy prediction in autonomous driving.
title MinkOcc: Towards real-time label-efficient semantic occupancy prediction
topic Computer Vision and Pattern Recognition
Robotics
url https://arxiv.org/abs/2504.02270