SAIL-Recon: Large SfM by Augmenting Scene Regression with Localization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Deng, Junyuan, Li, Heng, Xie, Tao, Ren, Weiqiang, Zhang, Qian, Tan, Ping, Guo, Xiaoyang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915462035537920
author Deng, Junyuan
Li, Heng
Xie, Tao
Ren, Weiqiang
Zhang, Qian
Tan, Ping
Guo, Xiaoyang
author_facet Deng, Junyuan
Li, Heng
Xie, Tao
Ren, Weiqiang
Zhang, Qian
Tan, Ping
Guo, Xiaoyang
contents Scene regression methods, such as VGGT, solve the Structure-from-Motion (SfM) problem by directly regressing camera poses and 3D scene structures from input images. They demonstrate impressive performance in handling images under extreme viewpoint changes. However, these methods struggle to handle a large number of input images. To address this problem, we introduce SAIL-Recon, a feed-forward Transformer for large scale SfM, by augmenting the scene regression network with visual localization capabilities. Specifically, our method first computes a neural scene representation from a subset of anchor images. The regression network is then fine-tuned to reconstruct all input images conditioned on this neural scene representation. Comprehensive experiments show that our method not only scales efficiently to large-scale scenes, but also achieves state-of-the-art results on both camera pose estimation and novel view synthesis benchmarks, including TUM-RGBD, CO3Dv2, and Tanks & Temples. We will publish our model and code. Code and models are publicly available at: https://hkust-sail.github.io/ sail-recon/.
format Preprint
id arxiv_https___arxiv_org_abs_2508_17972
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SAIL-Recon: Large SfM by Augmenting Scene Regression with Localization
Deng, Junyuan
Li, Heng
Xie, Tao
Ren, Weiqiang
Zhang, Qian
Tan, Ping
Guo, Xiaoyang
Computer Vision and Pattern Recognition
Scene regression methods, such as VGGT, solve the Structure-from-Motion (SfM) problem by directly regressing camera poses and 3D scene structures from input images. They demonstrate impressive performance in handling images under extreme viewpoint changes. However, these methods struggle to handle a large number of input images. To address this problem, we introduce SAIL-Recon, a feed-forward Transformer for large scale SfM, by augmenting the scene regression network with visual localization capabilities. Specifically, our method first computes a neural scene representation from a subset of anchor images. The regression network is then fine-tuned to reconstruct all input images conditioned on this neural scene representation. Comprehensive experiments show that our method not only scales efficiently to large-scale scenes, but also achieves state-of-the-art results on both camera pose estimation and novel view synthesis benchmarks, including TUM-RGBD, CO3Dv2, and Tanks & Temples. We will publish our model and code. Code and models are publicly available at: https://hkust-sail.github.io/ sail-recon/.
title SAIL-Recon: Large SfM by Augmenting Scene Regression with Localization
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2508.17972