OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Nextface

NextFace is a self-hosted PyTorch library that reconstructs 3D faces from one or more monocular RGB images, optimizing a statistical morphable model with autograd and ray tracing.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Nextface — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use NextFace when you can install a local Python environment and provide the required face-model files, and want to estimate a 3D face and related scene attributes from monocular RGB images.

Who it is for

  • Graphics developers who need facial geometry and appearance estimates from portrait images.
  • Researchers exploring inverse rendering or comparing reconstructions across images.
  • Developers prototyping pipelines that estimate face shape, reflectance, pose, camera parameters and illumination.

Documented uses

  • Estimate a 3D face from one RGB image.
  • Process multiple images in a batch.
  • Fit images of the same person with shared identity and reflectance attributes.
  • Estimate facial reflectance, head pose, camera parameters and scene lighting alongside geometry.
Before you choose

Requirements & limits

The library is self-hosted and licensed under GPL-3.0. Running it requires local setup and separately obtained face-model files.

  • The repository documents both CPU and CUDA-enabled GPU use.
  • The Basel Face Model is a separate prerequisite downloaded through its provider's form.
  • The repository provides an environment.yml and reports testing with torch 1.3.1, torchvision 0.4.2 and CUDA toolkit 10.1.
  • The optimizer supports a sharedIdentity mode for multiple images of the same person.

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

Reconstruct one person from multiple images

Use the documented shared-identity mode for several images of the same person.

Before you start

  • Complete the repository installation and model-file setup.
  • Put the person's images in one folder.

Steps

  1. Place the images for the same person in a folder.

  2. Run python optimizer.py --sharedIdentity --input path-to-your-folder-that-contains-all-your-images --output output-path-where-to-save-results.

Expected outcome

NextFace optimizes the images with shape identity and face reflectance shared across them.

Guide limitations

  • The repository's batch guidance specifies images with the same resolution.
  • The project describes improved geometry and reflectance estimation as a general tendency, not a guaranteed result.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
6 November 2026
Original website title
abdallahdib/NextFace · Official homepage title · retained in its source language
How we review sourcesSuggest a correction