Saved in:
Bibliographic Details
Main Authors: Kang, Gyeongjin, Yang, Seungkwon, Nam, Seungtae, Lee, Younggeun, Kim, Jungwoo, Park, Eunbyung
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.07806
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913175358668800
author Kang, Gyeongjin
Yang, Seungkwon
Nam, Seungtae
Lee, Younggeun
Kim, Jungwoo
Park, Eunbyung
author_facet Kang, Gyeongjin
Yang, Seungkwon
Nam, Seungtae
Lee, Younggeun
Kim, Jungwoo
Park, Eunbyung
contents We propose Multi-view Pyramid Transformer (MVP), a scalable multi-view transformer architecture that directly reconstructs large 3D scenes from tens to hundreds of images in a single forward pass. Drawing on the idea of ``looking broader to see the whole, looking finer to see the details," MVP is built on two core design principles: 1) a local-to-global inter-view hierarchy that gradually broadens the model's perspective from local views to groups and ultimately the full scene, and 2) a fine-to-coarse intra-view hierarchy that starts from detailed spatial representations and progressively aggregates them into compact, information-dense tokens. This dual hierarchy achieves both computational efficiency and representational richness, enabling fast reconstruction of large and complex scenes. We validate MVP on diverse datasets and show that, when coupled with 3D Gaussian Splatting as the underlying 3D representation, it achieves state-of-the-art generalizable reconstruction quality while maintaining high efficiency and scalability across a wide range of view configurations.
format Preprint
id arxiv_https___arxiv_org_abs_2512_07806
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Multi-view Pyramid Transformer: Look Coarser to See Broader
Kang, Gyeongjin
Yang, Seungkwon
Nam, Seungtae
Lee, Younggeun
Kim, Jungwoo
Park, Eunbyung
Computer Vision and Pattern Recognition
We propose Multi-view Pyramid Transformer (MVP), a scalable multi-view transformer architecture that directly reconstructs large 3D scenes from tens to hundreds of images in a single forward pass. Drawing on the idea of ``looking broader to see the whole, looking finer to see the details," MVP is built on two core design principles: 1) a local-to-global inter-view hierarchy that gradually broadens the model's perspective from local views to groups and ultimately the full scene, and 2) a fine-to-coarse intra-view hierarchy that starts from detailed spatial representations and progressively aggregates them into compact, information-dense tokens. This dual hierarchy achieves both computational efficiency and representational richness, enabling fast reconstruction of large and complex scenes. We validate MVP on diverse datasets and show that, when coupled with 3D Gaussian Splatting as the underlying 3D representation, it achieves state-of-the-art generalizable reconstruction quality while maintaining high efficiency and scalability across a wide range of view configurations.
title Multi-view Pyramid Transformer: Look Coarser to See Broader
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.07806