Compensating Motion-Induced Errors in Smartphone-Based VR Avatar Reconstruction

Recent developments in smartphone-based avatar reconstruction have made the creation of personalized and realistic avatars significantly more accessible. However, relying on one smartphone camera leads to capturing images sequentially, which introduces new challenges; particularly longer capture times increase the susceptibility to subject motion, which results in degraded reconstructions. We present a novel approach for smartphone-based avatar reconstruction that combines photogrammetry, silhouette constraints, and inverse rendering to produce high-fidelity, realistic avatars free of motion-induced artifacts. By using short, motion-resilient image sequences, referred to as sub-scans, we considerably reduce motion-induced artifacts. Our pipeline achieves high visual quality while offering improved robustness and outperforms current state-of-the-art methods in terms of computation time and accuracy.