A toolkit for creative video processing based on ML

Based on estimated depth and foreground (people), each video/webcam frame is represented as a 3D point cloud in real-time

Documentation

Features:

• Control point cloud depth scale, point size (base + variation), focus

• Add dissolve and chaos effects to create great dynamics

• Colorful outline for foreground which can be fully customized

• Work on background/foreground/outline layers (visibility/rays/desaturate)

• Separate examples based on compute shaders and Visual Effect Graph

• Rays/Gradient/Projector effects included

• A separate scene with Matrix-like effect for foreground (customize with your own design)

• A separate scene with video frames pages flow effect

• Several environments included (dynamic & static skyboxes)

• Lighting & post-process examples included (audio-sphere light, spot lights)

• MIDI control using GroundKontrol for VJ/show

Ideal for

❤️ Video processing

❤️ Media installations

❤️ Shows

❤️ VJing

Sample scenes included:

1. Depth point cloud (Shader graph / compute shader based)

2. Depth point cloud (VFX graph based)

3. Depth point pages

4. VFX Matrix glyphs

Bonus: a 5-star Awesome ML Kit asset is included for free and used as a base for ML estimation

Requirements:

• Unity 2021+, URP

• Windows 10/11

• Minimum Nvidia GTX 1050 or similar

• Recommended Nvidia RTX 3060 or similar