DuoSpaceNet: Leveraging Both Bird's-Eye-View and Perspective View Representations for 3D Object Detection

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Huang, Zhe, Zhao, Yizhe, Xiao, Hao, Wu, Chenyan, Ge, Lingting
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908305630167040
author Huang, Zhe
Zhao, Yizhe
Xiao, Hao
Wu, Chenyan
Ge, Lingting
author_facet Huang, Zhe
Zhao, Yizhe
Xiao, Hao
Wu, Chenyan
Ge, Lingting
contents Multi-view camera-only 3D object detection largely follows two primary paradigms: exploiting bird's-eye-view (BEV) representations or focusing on perspective-view (PV) features, each with distinct advantages. Although several recent approaches explore combining BEV and PV, many rely on partial fusion or maintain separate detection heads. In this paper, we propose DuoSpaceNet, a novel framework that fully unifies BEV and PV feature spaces within a single detection pipeline for comprehensive 3D perception. Our design includes a decoder to integrate BEV and PV features into unified detection queries, as well as a feature enhancement strategy that enriches different feature representations. In addition, DuoSpaceNet can be extended to handle multi-frame inputs, enabling more robust temporal analysis. Extensive experiments on nuScenes dataset show that DuoSpaceNet surpasses both BEV-based baselines (e.g., BEVFormer) and PV-based baselines (e.g., Sparse4D) in 3D object detection and BEV map segmentation, verifying the effectiveness of our proposed design.
format Preprint
id arxiv_https___arxiv_org_abs_2405_10577
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle DuoSpaceNet: Leveraging Both Bird's-Eye-View and Perspective View Representations for 3D Object Detection
Huang, Zhe
Zhao, Yizhe
Xiao, Hao
Wu, Chenyan
Ge, Lingting
Computer Vision and Pattern Recognition
Robotics
Multi-view camera-only 3D object detection largely follows two primary paradigms: exploiting bird's-eye-view (BEV) representations or focusing on perspective-view (PV) features, each with distinct advantages. Although several recent approaches explore combining BEV and PV, many rely on partial fusion or maintain separate detection heads. In this paper, we propose DuoSpaceNet, a novel framework that fully unifies BEV and PV feature spaces within a single detection pipeline for comprehensive 3D perception. Our design includes a decoder to integrate BEV and PV features into unified detection queries, as well as a feature enhancement strategy that enriches different feature representations. In addition, DuoSpaceNet can be extended to handle multi-frame inputs, enabling more robust temporal analysis. Extensive experiments on nuScenes dataset show that DuoSpaceNet surpasses both BEV-based baselines (e.g., BEVFormer) and PV-based baselines (e.g., Sparse4D) in 3D object detection and BEV map segmentation, verifying the effectiveness of our proposed design.
title DuoSpaceNet: Leveraging Both Bird's-Eye-View and Perspective View Representations for 3D Object Detection
topic Computer Vision and Pattern Recognition
Robotics
url https://arxiv.org/abs/2405.10577