Improving Distant 3D Object Detection Using 2D Box Supervision

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yang, Zetong, Yu, Zhiding, Choy, Chris, Wang, Renhao, Anandkumar, Anima, Alvarez, Jose M.
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916159669927936
author Yang, Zetong
Yu, Zhiding
Choy, Chris
Wang, Renhao
Anandkumar, Anima
Alvarez, Jose M.
author_facet Yang, Zetong
Yu, Zhiding
Choy, Chris
Wang, Renhao
Anandkumar, Anima
Alvarez, Jose M.
contents Improving the detection of distant 3d objects is an important yet challenging task. For camera-based 3D perception, the annotation of 3d bounding relies heavily on LiDAR for accurate depth information. As such, the distance of annotation is often limited due to the sparsity of LiDAR points on distant objects, which hampers the capability of existing detectors for long-range scenarios. We address this challenge by considering only 2D box supervision for distant objects since they are easy to annotate. We propose LR3D, a framework that learns to recover the missing depth of distant objects. LR3D adopts an implicit projection head to learn the generation of mapping between 2D boxes and depth using the 3D supervision on close objects. This mapping allows the depth estimation of distant objects conditioned on their 2D boxes, making long-range 3D detection with 2D supervision feasible. Experiments show that without distant 3D annotations, LR3D allows camera-based methods to detect distant objects (over 200m) with comparable accuracy to full 3D supervision. Our framework is general, and could widely benefit 3D detection methods to a large extent.
format Preprint
id arxiv_https___arxiv_org_abs_2403_09230
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Improving Distant 3D Object Detection Using 2D Box Supervision
Yang, Zetong
Yu, Zhiding
Choy, Chris
Wang, Renhao
Anandkumar, Anima
Alvarez, Jose M.
Computer Vision and Pattern Recognition
Improving the detection of distant 3d objects is an important yet challenging task. For camera-based 3D perception, the annotation of 3d bounding relies heavily on LiDAR for accurate depth information. As such, the distance of annotation is often limited due to the sparsity of LiDAR points on distant objects, which hampers the capability of existing detectors for long-range scenarios. We address this challenge by considering only 2D box supervision for distant objects since they are easy to annotate. We propose LR3D, a framework that learns to recover the missing depth of distant objects. LR3D adopts an implicit projection head to learn the generation of mapping between 2D boxes and depth using the 3D supervision on close objects. This mapping allows the depth estimation of distant objects conditioned on their 2D boxes, making long-range 3D detection with 2D supervision feasible. Experiments show that without distant 3D annotations, LR3D allows camera-based methods to detect distant objects (over 200m) with comparable accuracy to full 3D supervision. Our framework is general, and could widely benefit 3D detection methods to a large extent.
title Improving Distant 3D Object Detection Using 2D Box Supervision
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2403.09230