Real-Time Neural Hair Denoising

2026-05-17

AI Summary

This work addresses the noise and loss of fine details in hair rendering caused by severely undersampled rasterized inputs. We propose a lightweight, real-time method to reconstruct a strand-based hair G-buffer by integrating neural spatial reconstruction, temporal accumulation, and tangent-guided positional refinement—a novel combination that efficiently recovers per-pixel hair coverage and tangent orientation while enabling physically based deferred shading. Evaluated across diverse static and dynamic hairstyles, our approach consistently outperforms both specialized hair denoising techniques and general-purpose super-resolution methods such as DLSS and FSR in reconstruction fidelity, all while maintaining real-time performance.

Abstract

We propose a lightweight real-time method for reconstructing strand-based hair G-Buffers from severely undersampled rasterized inputs. Our pipeline first applies neural spatial reconstruction and temporal accumulation to recover hair coverage, i.e., fractional hair visibility within a pixel, and tangent. It then uses a tangent-guided reconstruction step to complete the position, which is subsequently used for physically based deferred hair shading. We evaluate our method across a diverse set of hairstyles, including straight, wavy, afro, and ponytail styles, under both static and dynamic scenarios. Our method achieves higher hair reconstruction quality than existing hair-specific denoising techniques and general industrial neural reconstruction solutions such as DLSS and FSR.

Problem

• hair reconstruction
• real-time rendering
• undersampled input
• G-Buffer
• hair denoising

Innovation

• neural hair denoising
• strand-based hair reconstruction
• real-time rendering
• tangent-guided reconstruction
• G-Buffer