Deep Inertial Prediction Challenge


The Deep Inertial Prediction (DIPr) challenge is a competition focused on improving short lasting IMU fallback tracking for VR systems.

DIPr (“dipper”) is a novel method for short-term IMU-only tracking for VR headsets that leverages deep learning which is intended to be used for fallback tracking for visual-inertial odometry or SLAM. To get started with the challenge you can visit the challenge website ( check out the instructions and sign up.