Mirror-Aware Neural Humans

TL;DR

This paper introduces a mirror-aware neural radiance field system that leverages single-camera mirror setups for accurate 3D human motion capture, including shape and dense appearance, by handling occlusions and calibrating automatically.

Contribution

It extends neural radiance fields to incorporate mirror information, enabling efficient 3D human body modeling from single-view data with automatic calibration and occlusion handling.

Findings

Improved 3D pose estimation accuracy in mirror scenes

Effective automatic calibration of camera and mirror orientation

Enhanced body modeling including shape and dense appearance

Abstract

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.