# TencentARC/Pixal3D

[SIGGRAPH 2026] Pixal3D: Pixel-Aligned 3D Generation from Images

Repository: https://github.com/TencentARC/Pixal3D
Canonical: https://ross.abutalabs.com/products/pixal3d
Homepage: https://ldyang694.github.io/projects/pixal3d/
Language: Python
License: MIT
License Family: permissive
Last push: 2026-06-23T10:10:27+00:00

## Health v2 (maintenance only)
Score: 54/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 89, release rhythm 35, longevity 8
- inputs: {"age_days": 115, "days_push": 71, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2156, forks 211 (observed 2026-08-28T04:06:19.719158+00:00)

## What it is
Pixal3D is a research codebase for generating high-fidelity 3D assets from a single image using a pixel-aligned generation paradigm that back-projects image features into 3D. It is a SIGGRAPH 2026 paper implementation from Tsinghua University and Tencent ARC Lab, built on the TRELLIS.2 backbone, with inference code, training code, and an online Gradio demo.

## Use cases
- generate a 3D model from a single photo
- create PBR-textured 3D assets for games from images
- reconstruct high-fidelity geometry matching an input image
- generate multi-view consistent 3D assets
- produce object-separated 3D scenes from images
- reproduce SIGGRAPH 2026 image-to-3D research results
- run image-to-3D generation in a browser demo

## When to choose
- you need pixel-level fidelity between the input image and the generated 3D asset
- you want detailed geometry with PBR textures from a single image
- you need multi-view or scene-level 3D generation from images
- you want to build on or study a state-of-the-art image-to-3D method with training code available

## When to avoid
- you need a lightweight tool without GPU requirements
- you want a production-ready end-user application rather than research code
- you need text-to-3D generation rather than image-to-3D
- you cannot install the TRELLIS.2 dependency stack

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, image-processing, graphics, simulation
- domain: computer-vision, graphics, artificial-intelligence
- platform: python
- tags: image-to-3d, 3d-generation, pixel-aligned, pbr-textures, siggraph-2026, gradio-demo, trellis, multi-view-generation, 3d-assets, research-code, game-development, gpu, linux, web-server

## Member repositories
- TencentARC/Pixal3D (main) score 54

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:19.719158+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:50:27.212134+00:00, confidence not recorded.
  - readme: https://github.com/TencentARC/Pixal3D (fetched 2026-08-28T04:06:19.719158+00:00, sha 93de2be0adaa)
  - homepage: https://ldyang694.github.io/projects/pixal3d/ (fetched 2026-08-29T10:30:30.407081+00:00, sha 5e532205d8ff)
- Data as of 2026-08-30T08:39:29.467469+00:00.