archive
Every paper Pith has read. Search by title, abstract, or pith.
456 papers in cs.GR · page 8
-
Explicit constraints preserve CAD design intent under edits
HistCAD: A Constraint-Aware Parametric History-Based CAD Representation, Dataset, and Benchmark with Industrial Complexity
-
Eye-tracking maps gaze to concepts in MOOC videos
COIVis: Eye-tracking-based Visual Exploration of Concept Learning in MOOC Videos
-
Matched 1D sections enforce curvature continuity in surface reconstruction
Thin Plate Spline Surface Reconstruction via the Method of Matched Sections
-
Transformer generates functional layouts by tokenizing building features
Tokenizing Buildings: A Transformer for Layout Synthesis
-
Sparse samples enable fast complex kernel convolutions
DISK: Differentiable Sparse Kernel Complex for Efficient Spatially-Variant Convolution
-
Pre-trained priors reuse motion data for any control task
SMP: Reusable Score-Matching Motion Priors for Physics-Based Character Control
-
Body reconstruction model regresses to average shapes
Investigating Anthropometric Fidelity in SAM 3D Body
-
Unified architecture handles 4D visualization and simulation
A Unified Architecture for N-Dimensional Visualization and Simulation: 4D Implementation and Evaluation including Boolean Operations
-
Deformation field tunes texture sampling density to visual frequency
FACT-GS: Frequency-Aligned Complexity-Aware Texture Reparameterization for 2D Gaussian Splatting
-
SURF delivers 12.5x faster 720p video generation
SURF: Signature-Retained Fast Video Generation
-
Shared MLP culls hidden Gaussians before 3D rendering
NVGS: Neural Visibility for Occlusion Culling in 3D Gaussian Splatting
-
Video cues align text-driven 3D motions with real dynamics
MotionDuet: Dual-Conditioned 3D Human Motion Generation with Video-Regularized Text Learning
-
Render gradients on deformation gradient reduce morphing silhouette error
PhysMorph-GS: Render-Guided Volumetric Morphing with Differentiable Physics
-
Scaling motion tracking yields generalist humanoid controller
SONIC: Supersizing Motion Tracking for Natural Humanoid Whole-Body Control
-
Gaussian split lifts dynamic 3D quality on sparse camera setups
Splatography: Sparse multi-view dynamic Gaussian Splatting for filmmaking challenges
-
Gesture realism saturates on BEAT2 while speech alignment claims weaken
Towards Reliable Human Evaluations in Gesture Generation: Insights from a Community-Driven State-of-the-Art Benchmark
-
Neural extraction recovers 3D point cloud watermarks at 83% accuracy
Deep Neural Watermarking for Robust Copyright Protection in 3D Point Clouds
-
One velocity field unifies three views of diffusion
The Principles of Diffusion Models
-
Synergy rewards plus time weighting fix RL conflicts in image generation
From Competition to Synergy: Unlocking Reinforcement Learning for Subject-Driven Image Generation
-
Attachment matching over motion assembles reusable 3D parts
Kinematic Kitbashing
-
Dual-path diffusion sharpens lip sync and head poses
KSDiff: Keyframe-Augmented Speech-Aware Dual-Path Diffusion for Facial Animation
-
Text Slider makes concept control 5x faster in diffusion models
Text Slider: Efficient and Plug-and-Play Continuous Concept Control for Image/Video Synthesis via LoRA Adapters
-
FFT reduces Bézier subdivision to O(d n log n) time
Fast subdivision of B\'{e}zier curves
-
New benchmark flags 12 types of misleading charts
Is this chart lying to me? Automating the detection of misleading visualizations
-
ViPE extracts 3D poses and depth from any raw video
ViPE: Video Pose Engine for 3D Geometric Perception
-
Hybrid solver handles sound radiation without ghost cells
SonicRadiation: A Hybrid Numerical Solution for Sound Radiation without Ghost Cells
-
Vertex features cut memory to one-fifth for neural illumination
Vertex Features for Neural Global Illumination
-
Neural BRDFs trained only on microgeometry random walks
PureSample: Neural Materials Learned by Sampling Microgeometry
-
Single video yields 3D smoke volumes via diffusion and physics
SmokeSVD: Smoke Reconstruction from A Single View via Progressive Novel View Synthesis and Refinement with Diffusion Models
-
Single GPU streams LoD Gaussians for city-scale scenes
A LoD of Gaussians: Unified Training and Rendering for Ultra-Large Scale Reconstruction with External Memory
-
Coarse NCA plus local decoder yields real-time high-res patterns
Neural Cellular Automata: From Cells to Pixels
-
B-splines give cloth simulation 2x speed with sharper wrinkles
Efficient B-Spline Finite Elements for Cloth Simulation
-
Generative frames fix monocular 4D reconstructions
BulletGen: Improving 4D Reconstruction with Bullet-Time Generation
-
Advection-diffusion PDE generalizes diffusion models
Beyond Blur: A Fluid Perspective on Generative Diffusion Models
-
Unified flow model preserves objects across repeated image edits
FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Space
-
Part graphs from language models synthesize realistic 4D interactions
HOI-PAGE: Zero-Shot Human-Object Interaction Generation with Part Affordance Guidance
-
Latent ODEs let 3D Gaussian Splatting extrapolate scenes into the future
ODE-GS: Latent ODEs for Dynamic Scene Extrapolation with 3D Gaussian Splatting
-
Pruned diffusion models recover quality with Jacobian matching
2ndMatch: Finetuning Pruned Diffusion Models via Second-Order Jacobian Matching
-
Benchmark adds particle trajectories to test physics in dynamic view synthesis
PhysGaia: A Physics-Aware Benchmark with Multi-Body Interactions for Dynamic Novel View Synthesis
-
Trajectories cut video tokens by 10x while raising accuracy
One Trajectory, One Token: Grounded Video Tokenization via Panoptic Sub-object Trajectory
-
LLMs turn text into valid CAD models using CadQuery scripts
CAD-Coder: Text-to-CAD Generation with Chain-of-Thought and Geometric Reward
-
Barrier optimization removes all penetration from fluid-shell and rod simulations
Penetration-free Solid-Fluid Interaction on Shells and Rods
-
Random linear map turns audio embeddings into dynamic visuals
LAV: Audio-Driven Dynamic Visual Generation with Neural Compression and StyleGAN2
-
New simulator merges Unreal graphics with precise MuJoCo physics
Unreal Robotics Lab: A High-Fidelity Robotics Simulator with Advanced Physics and Rendering
-
Text prompts create full 3D human textures via front-back image pairs
SMPL-GPTexture: Dual-View 3D Human Texture Estimation using Text-to-Image Generation Models
-
Containment queries on trimmed NURBS surfaces use a boundary formulation of the solid…
Robust Containment Queries over Collections of Trimmed NURBS Surfaces via Generalized Winding Numbers
-
Four families organize AI 3D radiological reconstructions
Representation Paradigms in AI-based 3D Radiological Image Reconstruction: A Systematic Review
-
Procedural hierarchy guides diffusion to edit building facades
Pro-DG: Procedural Diffusion Guidance for Architectural Facade Generation
-
Visual extraction and exit cues lift crowd model across scenarios
Improved visual-information-driven model for crowd simulation and its modular application
-
Real-time GI pipeline runs for dynamic 3D Gaussians and meshes
Real-time Global Illumination for Dynamic 3D Gaussian Scenes