Half-edge data structure

Half-edge data structure
Name	Half-edge data structure
Type	Data structure
Used in	Computational geometry, Computer graphics

Half-edge data structure

The half-edge data structure is a mesh representation used in computational geometry and computer graphics to encode the topology of polygonal meshes. It provides explicit connectivity by splitting each edge into two oppositely oriented half-edges, enabling efficient traversal and modification of surfaces, and is widely applied alongside algorithms from discrete differential geometry and numerical simulation.

Overview

The half-edge approach was popularized in the context of polygonal mesh processing by researchers influenced by early work from institutions such as IBM Research, Stanford University, Massachusetts Institute of Technology, and University of Utah. It became a core technique in software systems like OpenMesh, CGAL, Maya (software), and engines developed by id Software and Epic Games. The representation is discussed in textbooks and monographs used at Courant Institute of Mathematical Sciences, California Institute of Technology, and in courses at ETH Zurich.

Data Model and Components

A half-edge model encodes topology through primitive records: vertex, half-edge, and face. Each half-edge stores pointers to its twin half-edge, next half-edge around a face, origin vertex, and incident face; vertices often store an outgoing half-edge, and faces store a representative half-edge. Implementations across projects at Microsoft Research, Adobe Systems, Sony Interactive Entertainment, and academic groups at Princeton University and University of Pennsylvania adopt variations tailored for triangle meshes, quad meshes, or general polygonal meshes. The structure supports manifold and non-manifold representations used in contexts such as SIGGRAPH papers, Eurographics proceedings, and datasets like those from Stanford 3D Scanning Repository.

Operations and Algorithms

Typical operations include edge flips, edge collapses, edge splits, face insertion, and vertex smoothing, often employed in remeshing and subdivision schemes derived from work by practitioners at Weta Digital, Industrial Light & Magic, and labs affiliated with Harvard University. Traversal algorithms exploit half-edge pointers to enumerate incident faces, incident vertices, and boundary loops with complexity guarantees used in implementations at NVIDIA and in research at Google Research. Topology-aware algorithms for mesh simplification and parameterization use half-edge primitives in papers presented at venues such as ACM SIGGRAPH, IEEE Visualization Conference, and Symposium on Geometry Processing.

Variants and Related Structures

Variants include winged-edge, quad-edge, corner-table, and directed-edge structures, each emphasizing different trade-offs championed historically at organizations like Bell Labs and institutions such as University of California, Berkeley and Cornell University. The quad-edge by Guibas–Stolfi is prominent in computational geometry curricula at Brown University and University of Illinois Urbana-Champaign, while the corner-table has been used in game development at Rockstar Games and middleware libraries from Havok. Non-manifold extensions and generalized meshes are discussed in research from Tokyo Institute of Technology and Imperial College London.

Applications

Applications span modeling, animation, simulation, and digital fabrication. Studios like Pixar and research groups at Los Alamos National Laboratory use half-edge meshes for subdivision surfaces, collision detection, and finite element meshing workflows. In medical imaging projects at Mayo Clinic and Johns Hopkins University, half-edge representations support surface reconstruction and morphometry. Urban modeling efforts at MIT Media Lab and visualization systems developed at NASA and European Space Agency also incorporate half-edge–based pipelines for terrain and CAD-like data.

Implementation Considerations

Implementers choose storage strategies influenced by performance work from Intel Corporation and memory-efficiency studies from Apple Inc., balancing pointer-based records versus compressed representations suitable for mobile platforms from ARM Holdings. Concurrency, serialization, and GPU-friendly layouts have been advanced by teams at Valve Corporation, Blizzard Entertainment, and the Khronos Group; considerations include cache coherence, index-based versus pointer-based references, and support for half-edge mutation operations while maintaining manifold invariants studied at University of Toronto and University College London.

Category:Data structures Category:Computer graphics Category:Computational geometry