Stereo-Inertial Poser: Towards Metric-Accurate Shape-Aware Motion Capture Using Sparse IMUs and a Single Stereo Camera

The Appendix contains details on our in-house wearable IMU device, the time synchronization mechanism, networks structures, and network training. It also contains some qualitative results on AIST++ and TotalCapture datasets, which are also available on this website.

DEMO Video

This video contains our live demo and detailed qualitative comparison with existing baseline methods. We invite four subjects (volunteers) to record the demo.

Qualitative Results

Qualitative results on AIST++ dataset.

Qualitative results on TotalCapture dataset.

Body shape estimation. The top row visualize the initial point cloud and default SMPL model. The bottom row visualize the SMPL models fitted onto the point cloud with the optimization process.