The birth of the NUNO fabric scanner
- Angel Light
- Aug 3, 2025
- 2 min read
When it comes to digitizing fashion, you could say that the three most important elements are clothing, fabric, and avatar. Since the release of the Disney BRDF principle, the standards for physically based rendering have been largely established, and I believe that standardized texture scanning is now essential for the digitization of the fashion industry, so I started developing a prototype.
The NunoX fabric scanner is now a powerful device for capturing high-quality texture maps of fabric surfaces, and has proven itself to meet the requirements of various textile industries.
Prototype Development:
Hardware:
The normal (tangent) map is configured to represent angles in the brightness range of 0 to 1 on the XY axis, so the device design emits light up, down, left, and right, imaging the subject's brightness using visible light. During the experimental stage, it was more convenient to have a specification that made it easy to adjust the luminance and angle of the LED light source, so we built a frame using Lego (laughs). It was actually more expensive than the wooden frame specification that was initially shipped as a product.
I chose the iPhone as my camera because it only had one specification, which meant better lens quality, more reliable consistency, and simplified calibration.
software:
Tone mapping allows for intuitive visual adjustment of brightness and saturation, so as of 2016 I developed it using Adobe's ExtendScript Developer Tools, writing scripts while making adjustments in Photoshop, but later switched to Python's opencv.
Since then, performance has been improved through repeated hardware improvements such as gloss filtering, color temperature correction, linear space inverse transformation, and transparency detection, as well as software improvements such as diffuse reflection/albedo conversion and specular reflection calculation.It has now been reborn as the current NunoX, equipped with advanced features such as AI textures and AI physical property prediction.
Official website:
