A 20 minute video from an iPhone is now all you need to build a high precision 3D robot brain for any object.
Building perception models for robotics usually takes days of data collection and manual labeling. This new system automates the entire pipeline using synthetic data and rapid on-device training. It allows a user to capture an object and deploy a functional 3D model in under half an hour. This collapses a specialized engineering task into a simple consumer app experience. It means that robots can now be quickly adapted to new environments by people with zero technical expertise. The barrier between digital models and physical robots has finally vanished.
FalconApp: Rapid iPhone Deployment of End-to-End Perception via Automatically Labeled Synthetic Data
arXiv · 2604.25949
Reliable perception for robotics depends on large-scale labeled data, yet real-world datasets rely on heavy manual annotation and are time-consuming to produce. We present FalconApp, an iPhone app with an end-to-end frontend-backend pipeline that turns a short handheld capture of a rigid object into a perception module for mask detection and 6-DoF pose estimation. Our core contribution is a rapid mobile deployment pipeline paired with a photorealistic auto-labeling workflow: from a user-captured