Skip to main content
GameDev.net gamedev.net

Scientific Visualization News

The latest Scientific Visualization coverage curated for game developers.

Constrained Program Generation for 3D Reaction Animation with a 0.8B Model

arXiv cs.GR details ChemXRG, a DSL pipeline that turns 3D chemical reaction animation into executable programs. The team pairs persistent atom IDs with bond, charge, and grouping operations, then uses …

Research arXiv cs.GR · 18 hours, 26 minutes ago
Uncertainty-Aware Jacobi Set Computation

A new method for computing Jacobi sets now propagates uncertainty analytically instead of relying on repeated sampling. For teams doing multi-field visualization, that means more stable topological features when scalar …

Research arXiv cs.GR · 1 month ago
Fixed-Budget Gaussian Volume Encoding with Structure-Aware Allocation

A new Gaussian-volume encoding method packs scalar fields into a fixed primitive budget, then refines those primitives directly against the source data without densification or pruning. In tests, it cut …

Research arXiv cs.GR · 1 month ago
ESVR: 3D Ellipsoid-based Sparse Volume Rendering via Structure-aware Primitive Learning and Per-primitive Ray Sampling

A new sparse volume rendering method, ESVR, pushes large-scale volumetric visualization toward real-time use with ellipsoid primitives instead of dense voxels. It learns directly from 3D data, supports interactive transfer-function …

Research arXiv cs.GR · 1 month, 2 weeks ago
A Query-Efficient Stochastic Volume Rendering Framework for Time-Varying Implicit Neural Volumes

A new rendering framework for time-varying implicit neural volumes reaches 30-40 FPS at 1024x1024 on an RTX 4090 while keeping timestep changes to about 1-2 ms. It replaces dense ray …

Research arXiv cs.GR · 1 month, 3 weeks ago
3D Gaussian Splatting for Scientific Particle Data Compression and Rendering

3D Gaussian splatting is moving beyond pretty demos into scientific particle workflows. A new ParticleGS pipeline compresses hundreds of millions of particles into a renderable representation tuned for image quality, …

Research arXiv cs.GR · 1 month, 3 weeks ago
Loom: Multi-Region Analysis of Spatial Transcriptomics with Local Neighborhoods and Global Trajectories

Loom is a new spatial transcriptomics visualization system built for multi-region analysis, combining local neighborhood views with global pseudo-temporal trajectories. It uses a novel glyph and computational pipeline to help …

Research arXiv cs.GR · 1 month, 3 weeks ago
EVOLVE: Efficient Learned Volume Compression with Variable-Rate Encoding on a Cross-Domain Database

A new learned volume-compression system called EVOLVE targets high compression ratios for scientific simulation data without the per-volume tuning cost of implicit neural methods. Built around an autoencoder and variable-rate …

Research arXiv cs.GR · 2 months ago
Lossless-INR: Lossless Volumetric Implicit Neural Representations

Lossless-INR turns volumetric neural representations into exact reconstruction instead of “close enough” compression. The method decomposes voxel values into bit-planes, then learns each bit as a binary classification problem. For …

Research arXiv cs.GR · 2 months ago
DebrisTracer: Reliable Tracking in Hypervelocity Impact Fast Imaging

A new tracking framework, DebrisTracer, tackles a nasty computer-vision problem: following thousands of debris fragments in hypervelocity impact footage. By blending topology tracking with domain physics, it improves mass and …

Research arXiv cs.GR · 2 months ago
May (A)I Beautify Your Visualization? Expert Judgments of Acceptable Aesthetic Alterations

A survey of 95 visualization experts found that acceptability hinges more on what an alteration means than on whether it touches presentation or data. Small aesthetic tweaks like lighting or …

Research arXiv cs.GR · 2 months, 2 weeks ago
Visualizing Lagrangian Heat Transport Paths and Density Structures in Unsteady Heat Transfer

A new visualization method tackles unsteady heat transfer by following thermal transport in a Lagrangian, particle-based way instead of just showing temperature fields. The paper argues this exposes coherent heat …

Research arXiv cs.GR · 2 months, 2 weeks ago
HiLSVA: Design and Evaluation of a Human-in-the-Loop Agentic System for Scientific Visualization

A new human-in-the-loop agentic system for scientific visualization puts control back in the hands of the analyst instead of letting an LLM run the whole workflow. HiLSVA combines planning, provenance …

Research arXiv cs.GR · 2 months, 3 weeks ago
MS-COOT: Comparing Morse-Smale Complexes with Co-Optimal Transport

A new paper proposes treating Morse-Smale complexes as hypergraphs and comparing them with a co-optimal transport distance. For developers working on scientific visualization or topology-heavy tools, the interesting part is …

Research arXiv cs.GR · 3 months, 1 week ago
Streami: An MPI Data-Parallel Library to Compute Field Lines on GPUs

A new open-source GPU library, Streami, wraps MPI-based field-line computation in a thin layer that can plug into existing HPC applications. For teams doing fluid-flow visualization or in-situ analysis, the …

Research arXiv cs.GR · 3 months, 2 weeks ago
Chat Modeling: Interaction-Enhanced Agent Framework for Visualizing Literature-Grounded Biological Structures

A new agent framework turns natural-language prompts, paper text, and scene context into structured 3D modeling operations for biological visualization. For teams building editor tooling, the interesting bit is the …

Research arXiv cs.GR · 4 months ago
Exploring LLM Agent Designs and Interaction Modalities for Scientific Visualization

A benchmark study of 15 scientific-visualization tasks finds that general-purpose coding agents succeed most often, but they pay for it with higher compute and more overhead. Domain-specific agents are steadier …

Research arXiv cs.GR · 4 months, 3 weeks ago
Machine Learning for Scientific Visualization: Ensemble Data Analysis

This dissertation introduces advanced deep learning techniques for analyzing spatio-temporal scientific data. By utilizing autoencoders and the FLINT model, it addresses challenges like high dimensionality and missing information, which are …

Research arXiv cs.GR · 9 months, 3 weeks ago