A scan, a skeleton, a living asset — the rigging step that used to cost a week now costs nothing.
Creating production-ready 3D characters for AR, VR, or interactive applications has always had a hard constraint: rigging. A 3D scan gives you geometry. Turning that geometry into something that can move — walk, gesture, breathe — requires a skilled technical artist to manually define a skeleton, paint blend weights, and test deformations. For a single character, that is 8–40 hours of specialist work. Most SMEs cannot staff that. Most agencies charge for it. The result is that rich 3D character experiences stay inside AAA game studios and major film productions — out of reach for the companies that could use them in training simulations, retail try-on, cultural digitization, or live events.
A retailer scans a clothing sample on a mannequin, uploads the mesh, and 8 minutes later has a fully rigged character that can wear the garment in a virtual fitting room — responsive to body shape, animated to show drape and movement. A museum digitizes a bronze figure and publishes it as an interactive AR exhibit the same week. A training company creates a simulation populated with accurate, animated representations of the people and equipment in their actual facility. The 3D artist's role shifts from technical execution to creative direction. The pipeline that took a team weeks now takes one person an afternoon.
AI auto-rigging systems (Tripo3D, RigNet, and idezo's custom pipeline) use neural networks trained on thousands of rigged meshes to predict skeleton topology, joint placement, and blend weight maps from unseen geometry. The model learns that a humanoid scan needs hip/spine/shoulder/wrist joints at consistent anatomical proportions — and generalises this to novel body types, clothing layers, and partial scans.
The pipeline is: (1) scan capture via photogrammetry (any calibrated multi-camera rig or structured light scanner), (2) mesh cleanup and remeshing to a suitable polycount, (3) AI skeleton prediction and weight generation, (4) validation pass against a standard animation test clip, (5) export to your target runtime — Unity, Unreal, WebGL, or AR frameworks (ARKit/ARCore).
For characters requiring facial animation, a secondary blend shape generation step adds expression controls without manual sculpting. The result is compatible with standard FACS-based animation systems.
idezo builds and deploys this pipeline end-to-end — hardware, software, and training data — calibrated for your specific subject matter and target platform.