Enregistré dans:
Détails bibliographiques
Auteurs principaux: Li, Zhiqi, Dong, Chengrui, Chen, Yiming, Huang, Zhangchi, Liu, Peidong
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2503.10286
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866910873868566528
author Li, Zhiqi
Dong, Chengrui
Chen, Yiming
Huang, Zhangchi
Liu, Peidong
author_facet Li, Zhiqi
Dong, Chengrui
Chen, Yiming
Huang, Zhangchi
Liu, Peidong
contents We present VicaSplat, a novel framework for joint 3D Gaussians reconstruction and camera pose estimation from a sequence of unposed video frames, which is a critical yet underexplored task in real-world 3D applications. The core of our method lies in a novel transformer-based network architecture. In particular, our model starts with an image encoder that maps each image to a list of visual tokens. All visual tokens are concatenated with additional inserted learnable camera tokens. The obtained tokens then fully communicate with each other within a tailored transformer decoder. The camera tokens causally aggregate features from visual tokens of different views, and further modulate them frame-wisely to inject view-dependent features. 3D Gaussian splats and camera pose parameters can then be estimated via different prediction heads. Experiments show that VicaSplat surpasses baseline methods for multi-view inputs, and achieves comparable performance to prior two-view approaches. Remarkably, VicaSplat also demonstrates exceptional cross-dataset generalization capability on the ScanNet benchmark, achieving superior performance without any fine-tuning. Project page: https://lizhiqi49.github.io/VicaSplat.
format Preprint
id arxiv_https___arxiv_org_abs_2503_10286
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle VicaSplat: A Single Run is All You Need for 3D Gaussian Splatting and Camera Estimation from Unposed Video Frames
Li, Zhiqi
Dong, Chengrui
Chen, Yiming
Huang, Zhangchi
Liu, Peidong
Computer Vision and Pattern Recognition
We present VicaSplat, a novel framework for joint 3D Gaussians reconstruction and camera pose estimation from a sequence of unposed video frames, which is a critical yet underexplored task in real-world 3D applications. The core of our method lies in a novel transformer-based network architecture. In particular, our model starts with an image encoder that maps each image to a list of visual tokens. All visual tokens are concatenated with additional inserted learnable camera tokens. The obtained tokens then fully communicate with each other within a tailored transformer decoder. The camera tokens causally aggregate features from visual tokens of different views, and further modulate them frame-wisely to inject view-dependent features. 3D Gaussian splats and camera pose parameters can then be estimated via different prediction heads. Experiments show that VicaSplat surpasses baseline methods for multi-view inputs, and achieves comparable performance to prior two-view approaches. Remarkably, VicaSplat also demonstrates exceptional cross-dataset generalization capability on the ScanNet benchmark, achieving superior performance without any fine-tuning. Project page: https://lizhiqi49.github.io/VicaSplat.
title VicaSplat: A Single Run is All You Need for 3D Gaussian Splatting and Camera Estimation from Unposed Video Frames
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.10286