🤖 AI Summary
To address identity distortion, fairness deficits, temporal inconsistency, and non-real-time performance in real-time virtual makeup—caused by the entanglement of translucent makeup with skin tone and identity features—this paper proposes a two-stage disentanglement framework. Stage I generates pseudo-labels via unsupervised k-means clustering and graphics-based rendering, and jointly optimizes an alpha-weighted reconstruction loss and a lip-color-specific loss to achieve high-fidelity separation of makeup masks from skin tone. Stage II employs a lightweight, graphics-driven rendering module to ensure temporal coherence and fine-grained detail preservation. The method achieves real-time inference (>30 FPS) across diverse poses, expressions, and skin tones. It significantly improves transparency estimation accuracy, color fidelity, and identity preservation, outperforming state-of-the-art methods in detail reconstruction, motion smoothness, and cross-skin-tone robustness.
📝 Abstract
We present a novel framework for real-time virtual makeup try-on that achieves high-fidelity, identity-preserving cosmetic transfer with robust temporal consistency. In live makeup transfer applications, it is critical to synthesize temporally coherent results that accurately replicate fine-grained makeup and preserve user's identity. However, existing methods often struggle to disentangle semitransparent cosmetics from skin tones and other identify features, causing identity shifts and raising fairness concerns. Furthermore, current methods lack real-time capabilities and fail to maintain temporal consistency, limiting practical adoption. To address these challenges, we decouple makeup transfer into two steps: transparent makeup mask extraction and graphics-based mask rendering. After the makeup extraction step, the makeup rendering can be performed in real time, enabling live makeup try-on. Our makeup extraction model trained on pseudo-ground-truth data generated via two complementary methods: a graphics-based rendering pipeline and an unsupervised k-means clustering approach. To further enhance transparency estimation and color fidelity, we propose specialized training objectives, including alpha-weighted reconstruction and lip color losses. Our method achieves robust makeup transfer across diverse poses, expressions, and skin tones while preserving temporal smoothness. Extensive experiments demonstrate that our approach outperforms existing baselines in capturing fine details, maintaining temporal stability, and preserving identity integrity.