Zeyu Huang 黄泽宇

I am currently working on autonomous driving development in NIO.

I received my PhD degree in Computer Science from Shenzhen University supervised by Prof. Ruizhen Hu, working in Visual Computing Research Center.
Before that I got my B.Eng. in Software Engineering from Shenzhen University.

I am interested in Computer Graphics, Computer Vision and Robotics, especially on applying deep learning to synthesize graphics contents.

Email  /  Google Scholar  /  Github

I have strong interests in graphics content synthesis. Specifically, my researching projects cover the following topics: 3D Reconstruction, Interaction Generation, Object Manipulation, Floorplan Generation, etc.

In this paper, we introduce a novel method called FRI-Net for 2D floorplan reconstruction from 3D point cloud.

In this paper, we introduce a new method for the task of interaction transfer.

A means to generate interactions between two 3D objects with a descriptive and robust interaction representation.

A novel shape encoding for learning neural field representation of shapes that is category-agnostic and generalizable amid significant shape variations.

A descriptive and robust interaction representation of object manipulations to facilitate imitation learning.

A learning framework for automated floorplan generation which combines generative modeling using deep neural networks and user-in-the-loop designs to enable human users to provide sparse design constraints.

Template adapted from Yizhi and Jon.