#moveai
**Invisible** is a real-time, markerless motion capture solution powered by real-time convolutional neural networks and running on an RCX 6000 architecture. The technology utilizes CUDA and can operate in real-time on an RTX 1030. Invisible is capable of tracking 29 key points on the human body, delivering high-fidelity motion capture data streams with a 200-millisecond delay. This data can be rendered in the Unreal Engine and has a variety of applications in XR workflows. These include creating shadows under objects, real-time retargeting to a character rig such as a meta human, and triggering particle and gesture-based effects. These capabilities can be used to compile different graphics and trigger particle effects based on user movement.