Learning to Estimate Single-View Volumetric Flow Motions without 3D Supervision
Open access
Date
2023Type
- Conference Paper
ETH Bibliography
yes
Altmetrics
Abstract
We address the challenging problem of jointly inferring the 3D flow and volumetric densities moving in a fluid from a monocular input video with a deep neural network. Despite the complexity of this task, we show that it is possible to train the corresponding networks without requiring any 3D ground truth for training. In the absence of ground truth data we can train our model with observations from real-world capture setups instead of relying on synthetic reconstructions. We make this unsupervised training approach possible by first generating an initial prototype volume which is then moved and transported over time without the need for volumetric supervision. Our approach relies purely on image-based losses, an adversarial discriminator network, and regularization. Our method can estimate long-term sequences in a stable manner, while achieving closely matching targets for inputs such as rising smoke plumes. Show more
Permanent link
https://doi.org/10.3929/ethz-b-000654257Publication status
publishedExternal links
Book title
The Eleventh International Conference on Learning Representations (ICLR 2023)Publisher
OpenReviewEvent
Subject
Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Fluid Dynamics (physics.flu-dyn); FOS: Computer and information sciences; FOS: Physical sciencesOrganisational unit
08736 - Solenthaler, Barbara (Tit.-Prof.) / Solenthaler, Barbara (Tit.-Prof.)
Related publications and datasets
Is new version of: https://doi.org/10.48550/arXiv.2302.14470
More
Show all metadata
ETH Bibliography
yes
Altmetrics