OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Image & Video AI

Materialpalette

Materialpalette is a Python research repository that implements a method for extracting material concepts from a real-world image and region masks, generating textures, and decomposing them into albedo, normal and roughness maps.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Materialpalette — OpenViST project profile in Image & Video AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when you want to investigate or build on a code-based workflow for deriving PBR material information from a single real-world image.

Who it is for

  • Researchers studying single-image material extraction.
  • Developers exploring or adapting image-to-texture and material-map workflows.

Documented uses

  • Extract material concepts from images paired with region masks.
  • Generate texture images from pretrained concepts or concepts learned from selected images.
  • Decompose generated textures into albedo, normal and roughness maps.
Before you choose

Requirements & limits

The repository is public and released under the MIT license. Its documented setup uses Conda and Python; decomposition additionally requires the pretrained model archive.

  • This is a research code repository, not a documented hosted image-conversion service.
  • The workflow uses user-provided masks alongside images for concept extraction.
  • The project documents texture generation at different resolutions and provides four pretrained concepts for a generation-only starting point.
  • A pretrained model archive is required for decomposition, but not for texture extraction.
  • The README lists tested versions: Python 3.10.8, PyTorch 1.13, diffusers 0.19.3, peft 0.5 and PyTorch Lightning 1.8.3.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install Material Palette and prepare for decomposition

Set up the repository’s Conda environment and, if needed, download the pretrained decomposition model.

Before you start

  • Git and Conda are available.
  • The repository was tested with Python 3.10.8, PyTorch 1.13, diffusers 0.19.3, peft 0.5, and PyTorch Lightning 1.8.3.

Steps

  1. Clone the repository with the command below.

    Documented command
    git clone https://github.com/astra-vision/MaterialPalette.git

  2. Create the environment using conda env create --verbose -f deps.yml.

  3. Activate it with conda activate matpal.

  4. For decomposition, download the pretrained archive with wget https://github.com/astra-vision/MaterialPalette/releases/download/weights/model.tar.gz and untar it.

Expected outcome

The `matpal` Conda environment is activated; the pretrained model archive is downloaded for decomposition when that stage is needed.

Guide limitations

  • The pretrained model is not required for texture extraction alone.
  • The documented setup does not specify an untar command or give execution commands for the extraction or decomposition pipeline.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
17 October 2026
Original website title
astra-vision/MaterialPalette · Official homepage title · retained in its source language
How we review sourcesSuggest a correction