Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

GlassesGB: Controllable 2D GAN-Based Eyewear Personalization for 3D Gaussian Blendshapes Head Avatars

Rui-Yang Ju, Jen-Shiun Chiang

TL;DR

GlassesGB addresses the gap in VR eyewear personalization by enabling fine-grained, user-driven customization beyond predefined templates and 2D edits. It fuses controllable 2D eyewear design via GlassesGAN with 3D Gaussian Blendshapes and Gaussian Splatting to render eyewear as true 3D geometry on head avatars from monocular video. The approach demonstrates 3D-aware eyewear synthesis with temporal stabilization and preserves head geometry across viewpoints, contributing a first-of-its-kind bridge between 2D generative editing and 3D avatar rendering. This has practical implications for immersive VR try-on experiences, with future work focusing on stereo consistency and latency for real-time deployment.

Abstract

Virtual try-on systems allow users to interactively try different products within VR scenarios. However, most existing VTON methods operate only on predefined eyewear templates and lack support for fine-grained, user-driven customization. While GlassesGAN enables personalized 2D eyewear design, its capability remains limited to 2D image generation. Motivated by the success of 3D Gaussian Blendshapes in head reconstruction, we integrate these two techniques and propose GlassesGB, a framework that supports customizable eyewear generation for 3D head avatars. GlassesGB effectively bridges 2D generative customization with 3D head avatar rendering, addressing the challenge in achieving personalized eyewear design for VR applications. The implementation code is available at https://ruiyangju.github.io/GlassesGB.

GlassesGB: Controllable 2D GAN-Based Eyewear Personalization for 3D Gaussian Blendshapes Head Avatars

TL;DR

GlassesGB addresses the gap in VR eyewear personalization by enabling fine-grained, user-driven customization beyond predefined templates and 2D edits. It fuses controllable 2D eyewear design via GlassesGAN with 3D Gaussian Blendshapes and Gaussian Splatting to render eyewear as true 3D geometry on head avatars from monocular video. The approach demonstrates 3D-aware eyewear synthesis with temporal stabilization and preserves head geometry across viewpoints, contributing a first-of-its-kind bridge between 2D generative editing and 3D avatar rendering. This has practical implications for immersive VR try-on experiences, with future work focusing on stereo consistency and latency for real-time deployment.

Abstract

Virtual try-on systems allow users to interactively try different products within VR scenarios. However, most existing VTON methods operate only on predefined eyewear templates and lack support for fine-grained, user-driven customization. While GlassesGAN enables personalized 2D eyewear design, its capability remains limited to 2D image generation. Motivated by the success of 3D Gaussian Blendshapes in head reconstruction, we integrate these two techniques and propose GlassesGB, a framework that supports customizable eyewear generation for 3D head avatars. GlassesGB effectively bridges 2D generative customization with 3D head avatar rendering, addressing the challenge in achieving personalized eyewear design for VR applications. The implementation code is available at https://ruiyangju.github.io/GlassesGB.
Paper Structure (4 sections, 3 equations, 2 figures, 1 table)

This paper contains 4 sections, 3 equations, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Method
  3. Experiments
  4. Conclusion

Figures (2)

  • Figure 1: Examples of Customized Eyewear: The top row shows four synthesized eyewear styles corresponding to IDs 1–4. The table below lists the adjustable parameters used to generate each style.
  • Figure 2: Depth Map Comparison: We present pseudo-colored depth maps to show geometric details in the eyewear region.