RealX3D: A Physically-Degraded 3D Benchmark for Multi-view Visual Restoration and Reconstruction

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Liu, Shuhong, Bao, Chenyu, Cui, Ziteng, Liu, Yun, Chu, Xuangeng, Gu, Lin, Conde, Marcos V., Umagami, Ryo, Hashimoto, Tomohiro, Hu, Zijian, Xu, Tianhan, Gan, Yuan, Kurose, Yusuke, Harada, Tatsuya
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915742044127232
author Liu, Shuhong
Bao, Chenyu
Cui, Ziteng
Liu, Yun
Chu, Xuangeng
Gu, Lin
Conde, Marcos V.
Umagami, Ryo
Hashimoto, Tomohiro
Hu, Zijian
Xu, Tianhan
Gan, Yuan
Kurose, Yusuke
Harada, Tatsuya
author_facet Liu, Shuhong
Bao, Chenyu
Cui, Ziteng
Liu, Yun
Chu, Xuangeng
Gu, Lin
Conde, Marcos V.
Umagami, Ryo
Hashimoto, Tomohiro
Hu, Zijian
Xu, Tianhan
Gan, Yuan
Kurose, Yusuke
Harada, Tatsuya
contents We introduce RealX3D, a real-capture benchmark for multi-view visual restoration and 3D reconstruction under diverse physical degradations. RealX3D groups corruptions into four families, including illumination, scattering, occlusion, and blurring, and captures each at multiple severity levels using a unified acquisition protocol that yields pixel-aligned LQ/GT views. Each scene includes high-resolution capture, RAW images, and dense laser scans, from which we derive world-scale meshes and metric depth. Benchmarking a broad range of optimization-based and feed-forward methods shows substantial degradation in reconstruction quality under physical corruptions, underscoring the fragility of current multi-view pipelines in real-world challenging environments.
format Preprint
id arxiv_https___arxiv_org_abs_2512_23437
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle RealX3D: A Physically-Degraded 3D Benchmark for Multi-view Visual Restoration and Reconstruction
Liu, Shuhong
Bao, Chenyu
Cui, Ziteng
Liu, Yun
Chu, Xuangeng
Gu, Lin
Conde, Marcos V.
Umagami, Ryo
Hashimoto, Tomohiro
Hu, Zijian
Xu, Tianhan
Gan, Yuan
Kurose, Yusuke
Harada, Tatsuya
Computer Vision and Pattern Recognition
Multimedia
We introduce RealX3D, a real-capture benchmark for multi-view visual restoration and 3D reconstruction under diverse physical degradations. RealX3D groups corruptions into four families, including illumination, scattering, occlusion, and blurring, and captures each at multiple severity levels using a unified acquisition protocol that yields pixel-aligned LQ/GT views. Each scene includes high-resolution capture, RAW images, and dense laser scans, from which we derive world-scale meshes and metric depth. Benchmarking a broad range of optimization-based and feed-forward methods shows substantial degradation in reconstruction quality under physical corruptions, underscoring the fragility of current multi-view pipelines in real-world challenging environments.
title RealX3D: A Physically-Degraded 3D Benchmark for Multi-view Visual Restoration and Reconstruction
topic Computer Vision and Pattern Recognition
Multimedia
url https://arxiv.org/abs/2512.23437