Towards End-to-End Neuromorphic Event-based 3D Object Reconstruction Without Physical Priors

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xu, Chuanzhi, Chen, Langyi, Chen, Haodong, Chung, Vera, Qu, Qiang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908677282201600
author Xu, Chuanzhi
Chen, Langyi
Chen, Haodong
Chung, Vera
Qu, Qiang
author_facet Xu, Chuanzhi
Chen, Langyi
Chen, Haodong
Chung, Vera
Qu, Qiang
contents Neuromorphic cameras, also known as event cameras, are asynchronous brightness-change sensors that can capture extremely fast motion without suffering from motion blur, making them particularly promising for 3D reconstruction in extreme environments. However, existing research on 3D reconstruction using monocular neuromorphic cameras is limited, and most of the methods rely on estimating physical priors and employ complex multi-step pipelines. In this work, we propose an end-to-end method for dense voxel 3D reconstruction using neuromorphic cameras that eliminates the need to estimate physical priors. Our method incorporates a novel event representation to enhance edge features, enabling the proposed feature-enhancement model to learn more effectively. Additionally, we introduced Optimal Binarization Threshold Selection Principle as a guideline for future related work, using the optimal reconstruction results achieved with threshold optimization as the benchmark. Our method achieves a 54.6% improvement in reconstruction accuracy compared to the baseline method.
format Preprint
id arxiv_https___arxiv_org_abs_2501_00741
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Towards End-to-End Neuromorphic Event-based 3D Object Reconstruction Without Physical Priors
Xu, Chuanzhi
Chen, Langyi
Chen, Haodong
Chung, Vera
Qu, Qiang
Computer Vision and Pattern Recognition
Artificial Intelligence
Neuromorphic cameras, also known as event cameras, are asynchronous brightness-change sensors that can capture extremely fast motion without suffering from motion blur, making them particularly promising for 3D reconstruction in extreme environments. However, existing research on 3D reconstruction using monocular neuromorphic cameras is limited, and most of the methods rely on estimating physical priors and employ complex multi-step pipelines. In this work, we propose an end-to-end method for dense voxel 3D reconstruction using neuromorphic cameras that eliminates the need to estimate physical priors. Our method incorporates a novel event representation to enhance edge features, enabling the proposed feature-enhancement model to learn more effectively. Additionally, we introduced Optimal Binarization Threshold Selection Principle as a guideline for future related work, using the optimal reconstruction results achieved with threshold optimization as the benchmark. Our method achieves a 54.6% improvement in reconstruction accuracy compared to the baseline method.
title Towards End-to-End Neuromorphic Event-based 3D Object Reconstruction Without Physical Priors
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2501.00741