BulletGen: Improving 4D Reconstruction with Bullet-Time Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Rozumny, Denis, Luiten, Jonathon, Khan, Numair, Schönberger, Johannes, Kontschieder, Peter
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917388106072064
author Rozumny, Denis
Luiten, Jonathon
Khan, Numair
Schönberger, Johannes
Kontschieder, Peter
author_facet Rozumny, Denis
Luiten, Jonathon
Khan, Numair
Schönberger, Johannes
Kontschieder, Peter
contents Transforming casually captured, monocular videos into fully immersive dynamic experiences is a highly ill-posed task, and comes with significant challenges, e.g., reconstructing unseen regions, and dealing with the ambiguity in monocular depth estimation. In this work we introduce BulletGen, an approach that takes advantage of generative models to correct errors and complete missing information in a Gaussian-based dynamic scene representation. This is done by aligning the output of a diffusion-based video generation model with the 4D reconstruction at a single frozen "bullet-time" step. The generated frames are then used to supervise the optimization of the 4D Gaussian model. Our method seamlessly blends generative content with both static and dynamic scene components, achieving state-of-the-art results on both novel-view synthesis, and 2D/3D tracking tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2506_18601
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle BulletGen: Improving 4D Reconstruction with Bullet-Time Generation
Rozumny, Denis
Luiten, Jonathon
Khan, Numair
Schönberger, Johannes
Kontschieder, Peter
Graphics
Artificial Intelligence
Computer Vision and Pattern Recognition
Machine Learning
Transforming casually captured, monocular videos into fully immersive dynamic experiences is a highly ill-posed task, and comes with significant challenges, e.g., reconstructing unseen regions, and dealing with the ambiguity in monocular depth estimation. In this work we introduce BulletGen, an approach that takes advantage of generative models to correct errors and complete missing information in a Gaussian-based dynamic scene representation. This is done by aligning the output of a diffusion-based video generation model with the 4D reconstruction at a single frozen "bullet-time" step. The generated frames are then used to supervise the optimization of the 4D Gaussian model. Our method seamlessly blends generative content with both static and dynamic scene components, achieving state-of-the-art results on both novel-view synthesis, and 2D/3D tracking tasks.
title BulletGen: Improving 4D Reconstruction with Bullet-Time Generation
topic Graphics
Artificial Intelligence
Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2506.18601