Mirror-Aware Neural Humans
This work provides an affordable and reliable solution for 3D human motion capture in urban environments, though it is incremental by extending existing neural radiance fields to incorporate mirrors.
The paper tackles the problem of 3D human motion capture from single-view input by leveraging mirrors to provide two views, addressing depth ambiguities and occlusions. It introduces a mirror-aware neural radiance field method that learns a complete body model, achieving a consumer-level system with automatic calibration and improved performance in challenging mirror scenes.
Human motion capture either requires multi-camera systems or is unreliable when using single-view input due to depth ambiguities. Meanwhile, mirrors are readily available in urban environments and form an affordable alternative by recording two views with only a single camera. However, the mirror setting poses the additional challenge of handling occlusions of real and mirror image. Going beyond existing mirror approaches for 3D human pose estimation, we utilize mirrors for learning a complete body model, including shape and dense appearance. Our main contributions are extending articulated neural radiance fields to include a notion of a mirror, making it sample-efficient over potential occlusion regions. Together, our contributions realize a consumer-level 3D motion capture system that starts from off-the-shelf 2D poses by automatically calibrating the camera, estimating mirror orientation, and subsequently lifting 2D keypoint detections to 3D skeleton pose that is used to condition the mirror-aware NeRF. We empirically demonstrate the benefit of learning a body model and accounting for occlusion in challenging mirror scenes.