ScalarFlow: a large-scale volumetric data set of real-world scalar transport flows for computer animation and machine learning

  title={ScalarFlow: a large-scale volumetric data set of real-world scalar transport flows for computer animation and machine learning},
  author={Marie-Lena Eckert and Kiwon Um and Nils Thuerey},
  journal={ACM Trans. Graph.},
In this paper, we present ScalarFlow, a first large-scale data set of reconstructions of real-world smoke plumes. We additionally propose a framework for accurate physics-based reconstructions from a small number of video streams. Central components of our algorithm are a novel estimation of unseen inflow regions and an efficient regularization scheme. Our data set includes a large number of complex and natural buoyancy-driven flows. The flows transition to turbulent flows and contain… 

Neural Flow Map Reconstruction

This work introduces a technique for learning implicit neural representations of time‐varying vector fields that are specifically optimized to reproduce flow map samples sparsely covering the spatiotemporal domain of the data.

Learning Similarity Metrics for Volumetric Simulations with Multiscale CNNs

A similarity model based on entropy is proposed, which allows for the creation of physically meaningful ground truth distances for the similarity assessment of scalar and vectorial data, produced from transport and motion-based simulations.

A Fluid Flow Data Set for Machine Learning and its Application to Neural Flow Map Interpolation

This paper constructs a large fluid flow data set and applies it to a deep learning problem in scientific visualization, improving the accuracy of flow map interpolations, allowing a more precise flow analysis at a reduced memory IO footprint.

A Rapid, End‐to‐end, Generative Model for Gaseous Phenomena from Limited Views

This paper develops a novel conditional generative model to rapidly reconstruct the temporal density and velocity fields of gaseous phenomena based on the sequence of two projection views and can greatly improve the reconstruction performance for smoke scenes.

Optimization for Fluid Simulation and Reconstruction of Real-World Flow Phenomena

Based on a powerful convex optimization technique, the fast Primal-Dual method, algorithms for multi- and single-view reconstructions of real-world fluid phenomena, for flexible fluid guiding, and for realistic interaction of fluids and solids are developed.

Physics informed neural fields for smoke reconstruction with sparse data

This work presents the first method to reconstruct dynamic fluid phenomena by leveraging the governing physics in an end-to-end optimization from a mere set of sparse video frames without taking lighting conditions, geometry information, or boundary conditions as input.

Global Transport for Fluid Reconstruction with Learned Self-Supervision

A novel method to reconstruct volumetric flows from sparse views via a global transport formulation that reconstructs its motion based on a single initial state and introduces a learned self-supervision that constrains observations from unseen angles.

Learning Similarity Metrics for Numerical Simulations

A neural network-based approach that computes a stable and generalizing metric (LSiM) to compare data from a variety of numerical simulation sources and evaluates the different methods on a large range of test data to demonstrate the robustness of LSiM.

Fast Neural Representations for Direct Volume Rendering

A novel design of scene representation networks using GPU tensor cores to integrate the reconstruction seamlessly into on‐chip raytracing kernels are proposed, and the quality and performance of this network are compared to alternative network‐ and non‐network‐based compression schemes.

Deep Learning-based Upscaling for In Situ Volume Visualization

The specific adaptations and extensions that have been proposed in visualization to realize upscaling tasks beyond color images are shown, and how these approaches can be employed for in situ visualization are discussed.



Data-driven synthesis of smoke flows with CNN-based feature descriptors

This work employs a descriptor learning approach to encode the similarity between fluid regions with differences in resolution and numerical viscosity, and uses convolutional neural networks to generate the descriptors from fluid data such as smoke density and flow velocity.

Data-driven fluid simulations using regression forests

This paper proposes a novel machine learning based approach, that formulates physics-based fluid simulation as a regression problem, estimating the acceleration of every particle for each frame, and designed a feature vector, directly modelling individual forces and constraints from the Navier-Stokes equations.

Time-resolved 3d capture of non-stationary gas flows

This paper presents the first time-resolved Schlieren tomography system for capturing full 3D, non-stationary gas flows on a dense volumetric grid, and derives a new solution for this reconstruction problem that lends itself to efficient algorithms that robustly work with relatively small numbers of cameras.

Coupled Fluid Density and Motion from Single Views

A novel strategy to infer density updates strongly coupled to previous and current estimates of the flow motion is proposed, which employs an accurate discretization and depth‐based regularizers to compute stable solutions.

tempoGAN: A Temporally Coherent, Volumetric GAN for Super-resolution Fluid Flow

This work represents a first approach to synthesize four-dimensional physics fields with neural networks based on a conditional generative adversarial network that is designed for the inference of three-dimensional volumetric data, and generates consistent and detailed results.

From Capture to Simulation – Connecting Forward and Inverse Problems in Fluids

This paper explores the connection between fluid capture, simulation and proximal methods, finding that the proximal operator constraining fluid velocities to be divergence-free is directly equivalent to the pressure-projection methods commonly used in incompressible flow solvers.

Deep Fluids: A Generative Network for Parameterized Fluid Simulations

The proposed generative model is optimized for fluids by a novel loss function that guarantees divergence‐free velocity fields at all times, thus enabling applications such as fast construction of simulations, interpolation of fluids with different parameters, time re‐sampling, latent space simulations, and compression of fluid simulation data.

Example-based turbulence style transfer

This paper shows that style transfer methods from image editing can be adapted to transfer the turbulent style of an existing fluid simulation onto a new one by extending example-based image synthesis methods to handle velocity fields using a combination of patch-based and optimization-based texture synthesis.

Accelerating Eulerian Fluid Simulation With Convolutional Networks

This work proposes a data-driven approach that leverages the approximation power of deep-learning with the precision of standard solvers to obtain fast and highly realistic simulations of the Navier-Stokes equations.

Wavelet turbulence for fluid simulation

A novel wavelet method for the simulation of fluids at high spatial resolution that uses the wavelet decomposition of a low-resolution simulation to determine the location and energy characteristics of missing high-frequency components and synthesizes these missing components using a novel incompressible turbulence function.