Texture / PBR / Material Generation¶
3D Generation
Open Source¶
-
MaTe: Images Are All You Need for Material Transfer via Diffusion Transformer Discovered by AI
2026
A streamlined diffusion framework for material transfer that eliminates textual guidance and reference networks, integrating input images at the token level via multi-modal attention in a shared latent space.
diffusionmaterial-transfertexturetransformer -
Material Apprentice: Reflecting Process Expertise in Procedural Material Generation Discovered by AI
2026
A framework that formulates procedural material generation as retrieval-time process reasoning over expert demonstrations, automatically generating editable Blender node graphs rather than baked textures.
blendermaterialprocedural -
ReLi3D Discovered by AI
2026
A unified end-to-end pipeline developed by Stability AI that reconstructs 3D geometry, spatially-varying PBR materials, and environment illumination from sparse multi-view images in under one second.
3dmaterialpbrreconstruction -
Texture Space Material Diffusion Discovered by AI
2026
A framework by NVIDIA that finetunes a video diffusion transformer to generate high-quality PBR materials for 3D objects directly in texture space, supporting text-guided generation, multi-view material generation, and material upscaling up to 8K.
diffusionmaterialpbrtexture -
Toward Richer Material Generation via Procedural Data Enhancement Discovered by AI
2026
A method that enhances simple PBR materials into more expressive ones by augmenting the single GGX specular lobe into multi-lobe representations, improving the quality of training data for generative material models.
data-augmentationmaterialpbr -
Towards Realistic 3D Emission Materials: Dataset, Baseline, and Evaluation for Emission Texture Generation Discovered by AI
2026
A baseline model and dataset (Objaverse-Emission) designed for emission texture generation, enabling synthesized 3D objects to faithfully reproduce self-illuminating emission materials.
emissionmaterialpbrtexture -
VideoNeuMat Discovered by AI
2026
A two-stage pipeline that extracts reusable neural material assets from video diffusion models by fine-tuning a video model and performing inverse rendering.
materialpbrtexturevideo-diffusion -
2025
Ubisoft La Forge's chain-of-rendering decomposition that estimates PBR materials from generated texture images, turning AI textures into game-ready material maps.
material-estimationpbrubisoft -
Chord: Chain of Rendering Decomposition for PBR Material Estimation Discovered by AI
2025
Developed by Ubisoft La Forge, Chord is a two-stage generate-and-estimate framework that synthesizes shaded, tileable texture images using a fine-tuned diffusion model, then sequentially decomposes them into high-quality PBR material channels (albedo, normal, roughness, metallic) using a single-step image-conditioned diffusion model.
diffusionmaterialpbrtexture -
HiMat: DiT-based Ultra-High Resolution SVBRDF Generation Discovered by AI
2025
A Diffusion Transformer (DiT) framework designed for ultra-high-resolution (up to 4K) spatially varying bidirectional reflectance distribution function (SVBRDF) generation, maintaining pixel-level alignment across multiple reflectance maps.
ditpbrsvbrdftexture -
IntrinsiX: High-Quality PBR Generation using Image Priors Discovered by AI
2025
A text-to-PBR generation method that predicts physically-based rendering maps (albedo, roughness, metallic, and normal maps) from text descriptions without baked-in lighting, leveraging strong image priors.
materialpbrtext-to-texture -
LumiTex Discovered by AI
2025
An end-to-end framework that generates high-fidelity PBR textures by leveraging illumination context to decompose materials from image prompts and complete seamless, view-consistent textures.
diffusionmaterialpbrtexture -
MatE Discovered by AI
2025
A method for high-fidelity PBR material extraction from a single image under unconstrained real-world conditions, utilizing geometric priors to perform rectification and reduce distortion.
materialpbrsvbrdftexture -
2025
Generates physically-based materials (albedo, roughness, metallic, normal) for arbitrary 3D objects under varying illumination, producing relightable PBR assets.
diffusionmaterialpbrrelighting -
2025
Multi-view consistent PBR material generation for 3D meshes, predicting albedo and material channels that stay coherent across viewpoints.
materialmulti-viewpbr -
MaterialMVP: Illumination-Invariant Material Generation via Multi-view PBR Diffusion Discovered by AI
2025
A one-stage multi-view PBR diffusion model that generates high-quality, illumination-invariant PBR textures (albedo, metallic, roughness, normal) for 3D meshes from text or image prompts.
diffusionmaterialpbrtexture -
MatLat Discovered by AI
2025
A generative framework for producing high-quality PBR textures on 3D meshes by learning a dedicated material latent space (MatLat) through targeted fine-tuning of pretrained latent image generative models. It avoids distribution shifts when encoding additional PBR channels, enabling accurate representation of rough, metallic, and mixed materials.
latent-spacematerialpbrtexture -
MatMart: Material Reconstruction of 3D Objects via Diffusion Discovered by AI
2025
MatMart is a two-stage material reconstruction framework for 3D objects that combines accurate material prediction from inputs with prior-guided material generation for unobserved views. It utilizes progressive inference alongside a diffusion model to achieve high-fidelity, relightable PBR material recovery.
diffusionmaterialpbrtexture -
MatPedia: A Universal Generative Foundation for High-Fidelity Material Synthesis Discovered by AI
2025
A foundation model for material synthesis built upon a joint RGB-PBR representation that compactly encodes materials into interdependent latents for RGB appearance and PBR physical properties.
foundation-modelmaterialpbrtexture -
MatSpray Discovered by AI
2025
A framework that fuses 2D material world knowledge from diffusion models into 3D Gaussian Splatting geometry to reconstruct relightable assets with spatially varying PBR parameters.
gaussian-splattingmaterialpbrtexture -
MeshGen Discovered by AI
2025
An advanced image-to-3D pipeline that generates high-quality 3D meshes with detailed geometry and physically based rendering (PBR) textures.
3dmeshpbrtexture -
MuMA: 3D PBR Texturing via Multi-Channel Multi-View Generation and Agentic Post-Processing Discovered by AI
2025
A 3D PBR texturing method that models shaded and albedo appearance channels through multi-channel multi-view generation, integrated with intrinsic decomposition and agentic post-processing to produce high-quality, physically plausible materials.
agenticpbrtexture -
MVPainter: Accurate and Detailed 3D Texture Generation via Multi-View Diffusion with Geometric Control Discovered by AI
2025
MVPainter is a 3D texture generation framework that uses multi-view diffusion with geometric control to produce high-fidelity, detailed textures. It extracts physically-based rendering (PBR) attributes from the generated views to produce PBR-compliant meshes suitable for real-world rendering applications.
3ddiffusionpbrtexture -
NaTex: Native Texture Generation
2025
Native texture generation that synthesizes textures directly in 3D rather than by fusing multi-view 2D images, reducing seams and view-inconsistency artifacts.
diffusionnative-3dtexture -
PBR3DGen: A VLM-guided Mesh Generation with High-Quality PBR Texture Discovered by AI
2025
A two-stage framework that utilizes vision-language models (VLMs) to guide multi-view diffusion, generating high-quality 3D meshes with physically-based rendering materials that accurately capture reflective and metallic properties.
meshpbrtexturevlm -
2025
A multi-view diffusion approach to mesh texture generation that improves cross-view consistency and detail for production texturing.
diffusionmulti-viewtexture -
2025
Generates mesh textures by treating multi-view renders as a video sequence and leveraging a video diffusion model for spatially and temporally consistent results.
multi-viewtexturevideo-diffusion -
VLMaterial: Procedural Material Generation with Large Vision-Language Models Discovered by AI
2025
A framework that leverages large vision-language models to generate procedural materials represented as functional node graphs (Python programs) from input images.
materialpbrprocedural -
2024
Generates high-quality 3D assets with PBR materials from text or images using a primitive-based latent diffusion representation.
diffusionpbrprimitive -
ComfyUI-TextureAlchemy Discovered by AI
2024
A complete PBR and texture processing toolkit for ComfyUI designed to extract, process, and enhance physically-based rendering textures from single images for game engines and 3D pipelines.
comfyuipbrtexture-generation -
Cosmos-Transfer1-DiffusionRenderer Discovered by AI
2024
A dedicated video relighting and de-lighting framework based on the NVIDIA Cosmos World Foundation Models, enabling high-quality physical material estimation and controllable video lighting manipulation.
diffusionmaterialpbrrelighting -
DreamMat Discovered by AI
2024
A framework for high-quality PBR material generation that uses geometry- and light-aware diffusion models to resolve incorrect material decomposition and eliminate baked-in shading effects in albedo maps.
diffusionmaterialpbrtexture -
InTeX: Interactive Text-to-texture Synthesis via Unified Depth-aware Inpainting Discovered by AI
2024
An interactive text-to-texture synthesis framework that utilizes a unified depth-aware inpainting model to generate consistent and high-quality textures on 3D meshes while offering user-friendly interactive editing.
inpaintinginteractivetexture -
Jointly Generating Multi-view Consistent PBR Textures using Collaborative Control Discovered by AI
2024
A collaborative control diffusion model for text-to-texture generation that directly models full PBR stacks (including albedo, normal, and roughness maps) in view space. This approach ensures multi-view consistency, allowing the generated PBR textures to be seamlessly fused onto 3D meshes.
collaborative-controldiffusionpbrtexture -
Material Anything: Generating Materials for Any 3D Object via Diffusion Discovered by AI
2024
A fully-automated, unified diffusion framework designed to generate physically-based materials for diverse 3D objects under varying lighting conditions, supporting texture-less, albedo-only, and scanned meshes.
diffusionmaterialpbrtexture -
MaterialPicker: Multi-Modal Material Generation with Diffusion Transformers Discovered by AI
2024
A multi-modal material generator leveraging a Diffusion Transformer (DiT) architecture to simplify the creation of high-quality SVBRDF materials. It can generate complete PBR material maps conditioned on text prompts, image crops, or photographs of distorted surfaces.
ditmaterialpbrsvbrdf -
MatFusion Discovered by AI
2024
A generative diffusion model designed for SVBRDF estimation from photographs under various lighting conditions, utilizing an unconditional SVBRDF diffusion backbone trained on synthetic material exemplars.
diffusionpbrsvbrdftexture -
MCMat: Multiview-Consistent and Physically Accurate PBR Material Generation Discovered by AI
2024
A multiview-consistent and physically accurate PBR material generation framework that leverages multi-view diffusion priors to produce high-quality albedo, metallic, and roughness maps for 3D meshes.
diffusionmultiviewpbrtexture -
Neural LightRig: Unlocking Accurate Object Normal and Material Estimation with Multi-Light Diffusion Discovered by AI
2024
A framework that estimates accurate surface normals and PBR material properties from a single image by generating multi-light images using a diffusion model, providing rich auxiliary illumination cues.
diffusionmaterialnormal-mappbr -
PBRFusion Discovered by AI
2024
A community-developed diffusion model and ComfyUI extension widely used in RTX Remix workflows to upscale textures and generate high-quality normal, roughness, and depth maps for game remastering.
pbrrtx-remixtexture -
StableMaterials: Enhancing Diversity in Material Generation via Semi-Supervised Learning Discovered by AI
2024
A diffusion-based model designed for generating photorealistic PBR materials from text or image prompts, integrating semi-supervised learning with Latent Diffusion Models (LDMs).
diffusionpbrtextureGitHub · arXiv · Hugging Face
-
2024
Generates mesh textures by synchronizing a multi-view 2D diffusion process across rendered views, producing seamless texture maps without per-asset training.
diffusionmulti-viewtexture -
TexGaussian Discovered by AI
2024
A novel method that uses octant-aligned 3D Gaussian Splatting for rapid, high-quality Physically Based Rendering (PBR) material generation from text or image inputs.
gaussian-splattingmaterialpbrtexture -
2024
A feed-forward diffusion model that generates high-resolution textures directly in UV space, enabling fast text- and image-guided texturing of 3D meshes.
diffusionfeed-forwardtextureuv -
TexPro Discovered by AI
2024
A text-guided PBR texturing method that generates high-fidelity material maps for 3D meshes using procedural material modeling, enabling physically-based rendering and relighting.
materialpbrtexture -
MatFuse Discovered by AI
2023
A controllable material generation framework that leverages diffusion models to synthesize and edit SVBRDF maps using multi-modal inputs such as color palettes, sketches, text, and images.
diffusionmaterialpbrsvbrdf -
Paint-it Discovered by AI
2023
A text-driven high-fidelity PBR texture map synthesis method for 3D meshes via neural re-parameterized texture optimization.
materialpbrtexture -
Paint3D: Paint Anything 3D with Lighting-Less Texture Diffusion Models Discovered by AI
2023
A coarse-to-fine generative framework that produces high-resolution, lighting-less, and diverse 2K UV texture maps for untextured 3D meshes conditioned on text or image inputs. It generates textures without embedded illumination, allowing them to be re-lighted or re-edited within modern graphics pipelines.
diffusionpbrtexture
Proprietary / Industry Tools¶
-
PATINA Discovered by AI
2026
A commercial cloud API and model by fal.ai that generates seamless, high-resolution PBR material maps (BaseColor, Normal, Roughness, Metallic, and Height) from text descriptions or single input images.
apimaterialpbrtexture -
Adobe Substance 3D Sampler Discovered by AI
2021
An industry-standard tool for 3D material authoring and photogrammetry that features AI-powered 'Image to Material' capabilities. It uses machine learning to automatically convert single real-world photographs into complete, high-quality PBR materials with accurate normal, roughness, metallic, and height maps.
image-to-materialmaterialpbrphotogrammetry