Skip to main content
GameDev.net gamedev.net

Technical Art News

The latest Technical Art coverage curated for game developers.

ShapeLib: Designing a library of programmatic 3D shape abstractions with Large Language Models

ACM Graphics details ShapeLib, a 3D shape abstraction library designed with Large Language Models. The work lands in ACM Transactions on Graphics, Volume 46, Issue 2, and points to a …

Research ACM Graphics · 1 day, 9 hours ago
Worst-Case Non-parametric Shape Optimization

ACM Graphics details a worst-case non-parametric shape optimization method in ACM Transactions on Graphics. The April 2027 paper focuses on robust shape design under uncertainty, a topic that matters for …

Research ACM Graphics · 1 day, 9 hours ago
I Made a Pen Plotter to Draw Maps of the Moon

Sebastian Lague shows a DIY pen plotter that turns planetary data into contour maps of the Moon, Mars, and Earth. The build covers the mechanics, motor control, and pen holder …

Video Sebastian Lague · 1 week, 2 days ago
AgenticCADedit: A Stateful, Tool-Mediated Agentic Approach to Multimodal 3D CAD Editing

arXiv cs.GR details AgenticCADedit, a stateful CAD-editing pipeline that lets multimodal requests be applied as small, verifiable actions instead of regenerating a full model each time. For game teams building …

Research arXiv cs.GR · 2 weeks, 1 day ago
PIONEERS Shine Their LA Spotlight on Don Greenberg at SIGGRAPH 2026

SIGGRAPH says Don Greenberg, 92, headlined the Pioneers featured talk in a hybrid session at SIGGRAPH 2026, putting a long-view spotlight on one of computer graphics' early figures. For developers, …

Blog SIGGRAPH · 2 weeks, 3 days ago
How to Write with 3D Text in Marmoset Toolbag

Marmoset says Toolbag 5.03 adds Text Mode to Vector Layers in Texture Projects, letting artists place editable text directly on vector shapes. For material and texture workflows, that means faster …

Official Marmoset · 3 weeks ago
Articulate That Object Part (ATOP): 3D Part Articulation from Text and via Motion Personalization

ATOP is a new 3D part-articulation method that lets creators drive object motion from text and personalize it with motion examples. For game teams, that points to faster iteration on …

Research ACM Graphics · 3 weeks, 5 days ago
DealMaTe: Multi-Dimensional Material Transfer via Diffusion Transformer

DealMaTe proposes a diffusion-transformer approach for multi-dimensional material transfer, aiming to move more than just color or texture between assets. Published in ACM Transactions on Graphics, the work points toward …

Research ACM Graphics · 3 weeks, 5 days ago
Efficient Offset Mesh Generation with Preserved Features and Boundaries

Researchers have published a new offset-mesh generation method that preserves sharp features and boundaries while staying efficient enough for practical geometry workflows. The work targets a long-standing pain point in …

Research ACM Graphics · 3 weeks, 5 days ago
DMP: Directable Motion Retargeting through Motion Paraphrasing

ACM Graphics details DMP, a motion-retargeting method that lets teams steer character movement through motion paraphrasing instead of hand-tuning every clip. For animation-heavy games, that could mean more controllable reuse …

Research ACM Graphics · 3 weeks, 5 days ago
UniMate: One Unified Model to Animate Diverse Skeletons

UniMate aims to remove a major bottleneck in character animation: motion generation for arbitrary rigs. The model takes a rigged 3D asset and a text prompt, then produces articulated motion …

Research arXiv cs.GR · 1 month ago
No Pixel Left Behind: Filling Gaps in Anime Colorization

Anime colorization still burns time on tiny missed gaps, and a new tool called GapFill targets that pain point directly. Tested with 13 professional colorists, it improved gap-filling performance and …

Research arXiv cs.GR · 1 month, 1 week ago
Figures as Programs: Recursive Generation of Editable Scientific Figures

A new multi-agent system called FigTree turns scientific figure creation into recursive SVG program construction, aiming to make diagrams editable instead of locked raster images. It breaks a figure into …

Research arXiv cs.GR · 1 month, 1 week ago
Domain-Varying 2D Green' s Functions for Cage-based Deformation

A new cage-based deformation method, Domain-Varying Green Coordinates, reframes 2D shape editing around the Green’s function domain itself. The approach unifies harmonic and Green coordinates and lets artists tune deformation …

Research arXiv cs.GR · 1 month, 1 week ago
What Will This Copper Look Like Later? Forecasting Surface Appearance and Rendering It as a PBR Material

A new copper-oxidation pipeline forecasts how a fixed surface will look 10 accelerated units ahead, then turns that prediction into PBR maps for albedo, normal, roughness, and metallic. The practical …

Research arXiv cs.GR · 1 month, 1 week ago
Procedura: Agentic 3D Modeling with Procedural Control

Procedura turns text prompts into editable 3D assemblies instead of opaque meshes, aiming squarely at hard-surface modeling. The system uses an LLM to write parametric programs with named parts, typed …

Research arXiv cs.GR · 1 month, 1 week ago
Animating Octaves - Procedural GPU Patterns

Catlike Coding Godot shows how folding time into a GPU hash makes procedural octaves animate in lockstep, but not smoothly. With temporal lacunarity set to 2, the second octave updates …

Blog Catlike Coding Godot · 1 month, 1 week ago
A Three-Parameter Binary Subdivision Scheme for Shape-Controlled Curve Design

A new 9-point binary subdivision scheme adds three independent control parameters for shape-controlled curve design. Built from weighted refinement rules inspired by 7-point Lagrange and B-spline schemes, it gives developers …

Research arXiv cs.GR · 1 month, 2 weeks ago
ExMesh++: From Multi-View Images to Relightable UV-PBR Mesh Assets via Topology-Adaptive Reconstruction and Decomposition

ExMesh++ turns multi-view photos into editable UV-PBR mesh assets, aiming to skip the usual mesh extraction and texture-baking detour. The staged pipeline adapts topology with vertex split/merge operations, then locks …

Research arXiv cs.GR · 1 month, 2 weeks ago
ViSculpt: Visual-Centric Agentic Geometry Editing

A new Blender-focused geometry editor uses multimodal agents to modify existing meshes in place instead of regenerating them from scratch. ViSculpt watches the viewport, reasons about the current asset, and …

Research arXiv cs.GR · 1 month, 2 weeks ago
Page 1 of 5 Next