This page highlights three representative works.
[1] X. Ying, X. Wang, and Y. He. Saddle Vertex Graph (SVG): A Novel Solution to the Discrete Geodesic Problem, ACM Transactions on Graphics (Proceedings of ACM SIGGRAPH Asia 2013), Vol. 32, No. 6, Article No. 170, 2013.
Introduced a graph-based formulation showing that long discrete geodesic paths on polyhedral surfaces can be decomposed into reusable local segments anchored at saddle vertices. This shifted discrete geodesics away from purely source-dependent propagation toward reusable intrinsic structures.
[2] Y. Adikusuma, Z. Fang, and Y. He. Fast Construction of Discrete Geodesic Graphs, ACM Transactions on Graphics, Vol. 39, No. 2, Article No. 14, 2020 (presented at ACM SIGGRAPH Asia '20).
Extended the graph-based viewpoint with efficient, accuracy-aware construction, making scalable intrinsic distance infrastructure practical on large and anisotropic meshes.
Live capture of DGG construction and geodesic distance computation at interactive rates using the constructed graph (FastDGG, ACM TOG 2020).
[3] F. Hou, C. Wang, W. Wang, H. Qin, C. Qian, and Y. He. Iterative Poisson Surface Reconstruction (iPSR) for Unoriented Points, ACM Transactions on Graphics (Proceedings of ACM SIGGRAPH '22), Vol. 41, No. 4, Article No. 128, 2022.
Introduced a unified formulation for reconstruction from unoriented points, treating orientation and surface recovery as coupled aspects of a single geometric inference problem rather than as a brittle two-stage pipeline.
Illustration of iPSR convergence on 2D and 3D inputs with randomly initialized normals. In each iteration, iPSR updates point normals from the surface reconstructed in the previous iteration and then generates a new surface with improved quality. It typically converges within 3–20 iterations for common 3D models.