$\unicode{x1F493}$Heartian: Physiology-Aware Relightable Gaussian Head Avatar

2026-09-22

AI Summary

This study addresses the limitation of 3D Gaussian-based head avatars in neglecting subtle skin color variations induced by heartbeats. To overcome this, we propose a physiology-aware modulation framework built upon 3D Gaussian Splatting, dual-Gaussian waveform modeling, and lightweight MLP-based spatial residual learning. This approach encodes remote photoplethysmography (rPPG) signals into the material properties of relightable avatars for the first time, enabling the controllable embedding of heartbeat-induced skin color changes. Experimental results demonstrate that the proposed framework preserves reconstruction quality with a negligible PSNR degradation of only 0.005 dB while fully maintaining physiological signal recoverability. Specifically, it achieves a mean absolute error as low as 0.29 bpm for heart rate estimation, confirming that the embedded signals remain effectively detectable even after rendering.

Abstract

Gaussian head avatars typically model intrinsic facial appearance as temporally static, omitting subtle cardiac-induced skin-color variation. We propose $\unicode{x1F493}$Heartian, a physiology-aware modulation framework that learns cardiac-cycle-dependent per-frame albedo modulation of facial skin-region Gaussians within a relightable head avatar to encode remote photoplethysmography (rPPG) signals. Using synchronized contact PPG supervision, $\unicode{x1F493}$Heartian models the prescribed cardiac waveform as the sum of two Gaussian functions and learns per-frame spatial residuals via a lightweight MLP. Across 152 stationary recordings from UBFC-rPPG, PURE, and MMPD, attribute-space recovery of the supplied signal achieves a pooled recording-level heart-rate MAE of 0.29 bpm and MAPE of 0.38%. The signals remain detectable after rendering by benchmark rPPG methods, with the best tested configuration - a motion-augmented TS-CAN decoder pretrained on UBFC-rPPG - recovering heart rate from the rendered MMPD avatars at 0.97 bpm MAE and 1.21% MAPE. Meanwhile, $\unicode{x1F493}$Heartian maintains reconstruction quality comparable to the baseline, with negligible average PSNR degradation of 0.005 dB. Overall, our work embeds recoverable rPPG signals as controllable material attributes to subject-specific Gaussian head avatars while retaining the reconstruction quality.

Problem

• Gaussian head avatar
• remote photoplethysmography
• cardiac-induced skin-color variation
• physiological signal encoding

Innovation

• Relightable Gaussian Head Avatar
• Remote Photoplethysmography (rPPG)
• Physiology-Aware Modulation
• Albedo Modulation
• Lightweight MLP