Saved in:
Bibliographic Details
Main Authors: Wang, Ruicheng, Xu, Sicheng, Dai, Cassie, Xiang, Jianfeng, Deng, Yu, Tong, Xin, Yang, Jiaolong
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2410.19115
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912327675150336
author Wang, Ruicheng
Xu, Sicheng
Dai, Cassie
Xiang, Jianfeng
Deng, Yu
Tong, Xin
Yang, Jiaolong
author_facet Wang, Ruicheng
Xu, Sicheng
Dai, Cassie
Xiang, Jianfeng
Deng, Yu
Tong, Xin
Yang, Jiaolong
contents We present MoGe, a powerful model for recovering 3D geometry from monocular open-domain images. Given a single image, our model directly predicts a 3D point map of the captured scene with an affine-invariant representation, which is agnostic to true global scale and shift. This new representation precludes ambiguous supervision in training and facilitate effective geometry learning. Furthermore, we propose a set of novel global and local geometry supervisions that empower the model to learn high-quality geometry. These include a robust, optimal, and efficient point cloud alignment solver for accurate global shape learning, and a multi-scale local geometry loss promoting precise local geometry supervision. We train our model on a large, mixed dataset and demonstrate its strong generalizability and high accuracy. In our comprehensive evaluation on diverse unseen datasets, our model significantly outperforms state-of-the-art methods across all tasks, including monocular estimation of 3D point map, depth map, and camera field of view. Code and models can be found on our project page.
format Preprint
id arxiv_https___arxiv_org_abs_2410_19115
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle MoGe: Unlocking Accurate Monocular Geometry Estimation for Open-Domain Images with Optimal Training Supervision
Wang, Ruicheng
Xu, Sicheng
Dai, Cassie
Xiang, Jianfeng
Deng, Yu
Tong, Xin
Yang, Jiaolong
Computer Vision and Pattern Recognition
We present MoGe, a powerful model for recovering 3D geometry from monocular open-domain images. Given a single image, our model directly predicts a 3D point map of the captured scene with an affine-invariant representation, which is agnostic to true global scale and shift. This new representation precludes ambiguous supervision in training and facilitate effective geometry learning. Furthermore, we propose a set of novel global and local geometry supervisions that empower the model to learn high-quality geometry. These include a robust, optimal, and efficient point cloud alignment solver for accurate global shape learning, and a multi-scale local geometry loss promoting precise local geometry supervision. We train our model on a large, mixed dataset and demonstrate its strong generalizability and high accuracy. In our comprehensive evaluation on diverse unseen datasets, our model significantly outperforms state-of-the-art methods across all tasks, including monocular estimation of 3D point map, depth map, and camera field of view. Code and models can be found on our project page.
title MoGe: Unlocking Accurate Monocular Geometry Estimation for Open-Domain Images with Optimal Training Supervision
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2410.19115