Analysis updated 2026-05-18
Generate parametric 3D human head variations inside Blender using sliders.
Explore an in-progress semantic mapping system that turns PCA components into recognizable facial controls.
Follow early development of a lightweight, NumPy-only head generation addon.
Support ongoing development of the addon's facial component mapping work.
| controllerarts/gnm-controller | 0-bingwu-0/live-interpreter | 010zx00x1/faresnipe | |
|---|---|---|---|
| Stars | 2 | 2 | 2 |
| Language | Python | Python | Python |
| Setup difficulty | moderate | moderate | easy |
| Complexity | 3/5 | 2/5 | 2/5 |
| Audience | designer | general | general |
Figures from each repo's GitHub metadata at analysis time.
The README does not include install steps yet, the project is only about 5 percent complete.
GNM Controller is an early stage add on for Blender, the free 3D modeling software, that lets someone generate and adjust a realistic human head shape using sliders instead of manually sculpting it. It is built on a model called GNM Head, and instead of mapping raw mathematical numbers directly to controls, it uses a system the README calls semantic mapping, which is meant to translate those underlying numbers into attributes that make more sense to a person adjusting them, such as recognizable facial features rather than abstract statistical values. The README emphasizes that the tool is built for speed and simplicity. It relies only on NumPy, a widely used numerical computing library, and deliberately avoids heavier machine learning frameworks such as TensorFlow or Keras, so that it can update the 3D head shown in Blender's viewport in real time as sliders move, without extra software weighing it down. According to the README, the project is very early: the author states it is at five percent completion. The basic engine can already generate different mesh variations, but the bulk of the remaining work is going through 253 individual mathematical components one by one to figure out what facial feature each one actually controls, and organizing that information into a stable, usable mapping without breaking the underlying 3D mesh structure. The README is short and does not include installation steps, usage instructions, or license information, so at this stage the project should be treated as an early, in progress addon rather than a finished, ready to install tool. The developer has set up a Ko-fi page to accept support for the ongoing work of building out this mapping system.
An early-stage Blender addon that lets you shape a realistic 3D human head using sliders instead of manual sculpting.
Mainly Python. The stack also includes Python, NumPy, Blender.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly designer.
This repo across BitVibe Labs
Verify against the repo before relying on details.