Xingyi Du (杜兴逸)

Xingyi Du (杜兴逸)

Researcher at Lightspeed Studios, Tencent America

Lightspeed Studios, Tencent America

About Me

I am a researcher at Lightspeed Studios, Tencent America. I obtained Ph.D. from CSE, Washington University in St. Louis, where I studied computer graphics and geometric modeling under the supervision of Prof. Tao Ju. My PhD thesis focuses on injective mapping of triangular and tetrahedral meshes under constraints such as fixed boundary or positional handles. I also explored modeling shapes with sharp features and complex structures using multiple implicit functions. I am fortunate to collaborate with brilliant researchers at Adobe Research on many exciting problems in graphics and geometry.

I earned my master’s degree from Tsinghua University in 2018. There, I worked on high quality remeshing with Prof. Hui Zhang and Prof. Dong-Ming Yan. I finished my undergraduate study in 2015 from Tsinghua University with a major in Material Science (Physics) and a minor degree in Computer Science.

Interests

  • Geometry Processing
  • Implicit Modeling
  • Optimization

Education

  • Ph.D. in Computer Science, 2023

    Washington University at St. Louis

  • M.Sc. in Software Engineering, 2018

    Tsinghua University

  • Minor in Computer Science, 2015

    Tsinghua University

  • B.E. in Material Science and Engineering, 2015

    Tsinghua University

Recent Publications

Lifted Surfacing of Generalized Sweep Volumes

A general algorithm for computing watertight sweep volume boundaries with sharp geometric features via lifting to higher dimensions

RL-ACD: Reinforcement Learning-based Approximate Convex Decomposition

A data-driven reinforcement learning approach for efficient and near-optimal approximate convex decomposition of 3D shapes

Adaptive Grid Generation for Discretizing Implicit Complexes

An adaptive grid generation method that captures fine geometric details for implicit surfaces including their intersection curves and joint points

Isometric Energies for Recovering Injectivity in Constrained Mapping

A method for computing injective (one-to-one) and low-distortion maps under fixed-boundary or positional constraints

Robust Computation of Implicit Surface Networks for Piecewise Linear Functions

A robust method for computing implicit surface networks (e.g. arrangements and material interfaces) on a tetrahedron grid

Recent Posts

如何顺利通过 PhD Oral Exam

上个月最后一个周五下午,我坐在电脑前,望着zoom的空白窗口,心里一块大石头高高悬着。在屏幕那一边,三位教授正讨论着。几分钟后,当我听到教授说出“congrats”时,心里的石头终于落地,我终于通过了口试。