基本情况
李天行,讲师,硕士生导师。2021年于日本东京大学取得博士学位,随后加入北京工业大学,从事教学科研工作。
围绕计算机图形学,开展几何三维重建、变形、动画等方面的研究。在该领域发表CCF推荐A/B类期刊/会议论文多篇,并多次受邀于国际会议发表。主持/参与国家自然科学基金、北京市自然科学基金等多项研究课题。
围绕人工智能、三维变形体仿真方向,指导多名本/硕生参加竞赛、课题,并形成高水平SCI论文及发明专利。
代表性研究成果
[1] T. Li, R. Shi, Q. Zhu, T. Kanai. SwinGar: Spectrum-Inspired Neural Dynamic Deformation for Free-Swinging Garments[J]. IEEE Transactions on Visualization and Computer Graphics, 2024, 30(10): 6913-6927. (Presented at Pacific Graphics 2024)
[2] T. Li, R. Shi, T. Kanai. MultiResGNet: Approximating Nonlinear Deformation via Multi-Resolution Graphs[J]. Computer Graphics Forum, 2021, 40(2): 537–548. (Presented at Eurographics 2021)
[3] T. Li, R. Shi, T. Kanai. Detail-Aware Deep Clothing Animations Infused with Multi-Source At-tributes[J]. Computer Graphics Forum, 2023, 42(1): 231-244. (Presented at Eurographics 2023)
[4] T. Li, R. Shi, T. Kanai. DenseGATs: A Graph-Attention-Based Network for Nonlinear Character Deformation[C]. ACM SIGGRAPH Symposium on Interactive 3D Graphics and Games (I3D), 2020.
[5] T. Li, R. Shi, Z. Li, T. Kanai, Q. Zhu. Efficient Deformation Learning of Varied Garments with a Structure-Preserving Multilevel Framework[C]. ACM SIGGRAPH Symposium on Interactive 3D Graphics and Games (I3D), 2024.