Guardado en:
Detalles Bibliográficos
Autores principales: Foucard, Louis, Khanna, Samar, Shi, Yi, Liu, Chi-Kuei, Shen, Quinn Z, Ngo, Thuyen, Xia, Zi-Xiang
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2405.15843
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911886926151680
author Foucard, Louis
Khanna, Samar
Shi, Yi
Liu, Chi-Kuei
Shen, Quinn Z
Ngo, Thuyen
Xia, Zi-Xiang
author_facet Foucard, Louis
Khanna, Samar
Shi, Yi
Liu, Chi-Kuei
Shen, Quinn Z
Ngo, Thuyen
Xia, Zi-Xiang
contents In this paper, we propose SpotNet: a fast, single stage, image-centric but LiDAR anchored approach for long range 3D object detection. We demonstrate that our approach to LiDAR/image sensor fusion, combined with the joint learning of 2D and 3D detection tasks, can lead to accurate 3D object detection with very sparse LiDAR support. Unlike more recent bird's-eye-view (BEV) sensor-fusion methods which scale with range $r$ as $O(r^2)$, SpotNet scales as $O(1)$ with range. We argue that such an architecture is ideally suited to leverage each sensor's strength, i.e. semantic understanding from images and accurate range finding from LiDAR data. Finally we show that anchoring detections on LiDAR points removes the need to regress distances, and so the architecture is able to transfer from 2MP to 8MP resolution images without re-training.
format Preprint
id arxiv_https___arxiv_org_abs_2405_15843
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SpotNet: An Image Centric, Lidar Anchored Approach To Long Range Perception
Foucard, Louis
Khanna, Samar
Shi, Yi
Liu, Chi-Kuei
Shen, Quinn Z
Ngo, Thuyen
Xia, Zi-Xiang
Computer Vision and Pattern Recognition
Artificial Intelligence
In this paper, we propose SpotNet: a fast, single stage, image-centric but LiDAR anchored approach for long range 3D object detection. We demonstrate that our approach to LiDAR/image sensor fusion, combined with the joint learning of 2D and 3D detection tasks, can lead to accurate 3D object detection with very sparse LiDAR support. Unlike more recent bird's-eye-view (BEV) sensor-fusion methods which scale with range $r$ as $O(r^2)$, SpotNet scales as $O(1)$ with range. We argue that such an architecture is ideally suited to leverage each sensor's strength, i.e. semantic understanding from images and accurate range finding from LiDAR data. Finally we show that anchoring detections on LiDAR points removes the need to regress distances, and so the architecture is able to transfer from 2MP to 8MP resolution images without re-training.
title SpotNet: An Image Centric, Lidar Anchored Approach To Long Range Perception
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2405.15843