MediaPipe Hands: On-device Real-time Hand Tracking
This enables on-device hand tracking for AR/VR users, but it is incremental as it builds on existing hand tracking methods with a new implementation.
The paper tackles real-time hand tracking from a single RGB camera for AR/VR applications by proposing a pipeline with a palm detector and hand landmark model, achieving real-time inference on mobile GPUs and high prediction quality.
We present a real-time on-device hand tracking pipeline that predicts hand skeleton from single RGB camera for AR/VR applications. The pipeline consists of two models: 1) a palm detector, 2) a hand landmark model. It's implemented via MediaPipe, a framework for building cross-platform ML solutions. The proposed model and pipeline architecture demonstrates real-time inference speed on mobile GPUs and high prediction quality. MediaPipe Hands is open sourced at https://mediapipe.dev.