Pre-computed Liquid Spaces with Generative Neural Networks and Optical Flow

By Boris Bonev, Lukas Prantl, Nils Thürey. 2017

15 citations recorded by the source corpus.

Liquids exhibit highly complex, non-linear behavior under changing simulation conditions such as user interactions. We propose a method to map this complex behavior over a parameter range onto a reduced representation based on space-time deformations. In order to represent the complexity of the full space of inputs, we use aligned deformations from optical flow solves, and we leverage the power of generative neural networks to synthesize additional deformations for refinement. We introduce a novel deformation-aware loss function, which enables optimization in the highly non-linear space of multiple deformations. To demonstrate the effectiveness of our approach, we showcase the method with several complex examples in two and four dimensions. Our representation makes it possible to generate implicit surfaces of liquids very efficiently, which allows us to very efficiently display the scene from any angle, and to add secondary effects such as particle systems. We have implemented a mobile application with our full pipeline to demonstrate that real-time interaction is possible with our approach.

Subfields: Computational Mechanics, Computer Graphics and Computer-Aided Design, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

In Computer Graphics and Computer-Aided Design

In Computer Vision and Pattern Recognition

Citations in the corpus

Latent Space Physics: Towards Learning the Temporal Evolution of Fluid Flow (2018, methodology): “, by using space-time deformations [Prantl et al. 2017; Raveendran et al. 2014], but we will focus on simulation algorithms that explicitly compute transport processes with advection algorithms in the following.”
Data-driven synthesis of smoke flows with CNN-based feature descriptors (2017, unknown): “…have demonstrated the inference of SPH forces with regression forests [Ladicky et al. 2015], proposed CNNs for fast pressure projections [Tompson et al. 2016], learned space-time deformations for interactive liquids [Prantl et al. 2017] , and modeled splash statistics with NNs [Um et al. 2017].”
Fluid directed rigid body control using deep reinforcement learning (2018, methodology): “Besides keyframe-based methods, there are other techniques [Bonev et al. 2017; Raveendran et al. 2014; Thuerey 2016] that synthesize fluid animations by interpolating existing animation data.”