Guoxian Song

Title: Guoxian Song: Innovator in 3D Image Synthesis

Introduction

Guoxian Song is a prominent inventor based in Los Angeles, CA, known for his contributions to the field of 3D image synthesis. With a total of five patents to his name, he has made significant strides in developing technologies that enhance the quality and control of synthesized images.

Latest Patents

One of Guoxian Song's latest patents is titled "Geometry-guided controllable 3D head synthesis." This innovative technology describes a framework for producing high-quality 3D head images with full control over various parameters, including camera poses, facial expressions, head shape, and articulated neck and jaw poses. The patent also introduces a semantic signed distance function (SDF) formulation that allows for volumetric correspondence from observation space to canonical space, enabling full disentanglement of control parameters during 3D GAN training. Another notable patent is "Cascaded domain bridging for image generation," which outlines a method for generating stylized 3D avatars. This method involves using a generative adversarial network (GAN) to create a stylized image based on an input image, followed by generating an avatar image through a trained model.

Career Highlights

Throughout his career, Guoxian Song has worked with several companies, including Lemon Inc. and Beijing Zitiao Network Technology Co., Ltd. His work has focused on advancing the capabilities of image synthesis and avatar generation, making him a key player in the tech industry.

Collaborations

Guoxian has collaborated with notable coworkers such as Jing Liu and Linjie Luo, contributing to the development of innovative technologies in his field.

Conclusion

Guoxian Song's work in 3D image synthesis and avatar generation showcases his innovative spirit and dedication to advancing technology. His patents reflect a commitment to enhancing the quality and control of synthesized images, making a significant impact in the industry.