Tinghui Zhou

Title: Tinghui Zhou and View-Consistent Texture Generation

Introduction

Tinghui Zhou is an inventor based in San Mateo, California, and is associated with Roblox Corporation. His patent work addresses the generation of consistent textures for three-dimensional objects using generative machine learning.

Latest Patents

Zhou is credited with 1 patent: “View Consistent Texture Generation for Three Dimensional Objects.” The invention describes methods, systems, and computer-readable media for creating textures that remain consistent across different views of a 3D object.

The method generates multiple depth maps from a 3D mesh, with each map corresponding to a view of the object. It combines those depth maps with a text prompt describing the desired texture and provides them to a generative machine-learning model, which produces two or more views of a texture map. The views are then combined using the 3D mesh to create the object’s texture map.

Career Highlights

Zhou’s listed patent connects generative machine learning with 3D modeling and texture creation. The approach is designed to help produce textures that cover a 3D mesh while maintaining consistency across multiple viewpoints.

Collaborations

Zhou’s listed coworkers include Maneesh Agrawala and Timothy Paul Omernick. His work is associated with Roblox Corporation.

Conclusion

Tinghui Zhou’s patent presents a machine-learning approach to generating view-consistent textures for 3D objects, combining depth information, text descriptions, and mesh-based processing.