Ramin Ebrahim Nakhli

Title: Ramin Ebrahim Nakhli: Inventor in Generative 3D Modeling.

Introduction

Ramin Ebrahim Nakhli is an inventor based in Thornhill, Canada, and works at Qualcomm Incorporated. His patent work concerns machine learning methods for generating three-dimensional meshes from text.

Latest Patents

Nakhli is associated with 1 patent: “Text-to-image diffusion models for generalizable mesh generation.” The disclosed method processes a textual input with a diffusion machine learning model to generate a multiview latent tensor representing multiple orthographic projections. A conversion model then uses that tensor to generate a triplane latent tensor, which a decoder model processes to produce a three-dimensional mesh.

Career Highlights

Nakhli’s patent describes a pipeline that connects text-to-image diffusion techniques with 3D mesh generation. The approach uses intermediate latent representations to translate textual descriptions into three-dimensional geometry.

Collaborations

The patent lists Antoine Clement Mercier and Guillaume Jean Fernand Berger among Nakhli’s coworkers.

Conclusion

Ramin Ebrahim Nakhli’s patent work at Qualcomm Incorporated focuses on using machine learning to generate three-dimensional meshes from textual inputs.