A mesh represents a surface as many simple connected pieces, most often triangles. Instead of solving a separate geometric problem for every kind of smooth surface, an algorithm can work with segments and triangles. That uniformity makes meshes practical for operations such as intersections, offsets, and Boolean modeling, where intersections become polylines rather than circles or other specialized curves. section 11.2.2 section 11.2.3
A triangle mesh commonly stores a vertex table plus faces that refer to vertex indices. Sharing vertices avoids repeating coordinates: a cube surface can use eight vertices and twelve triangular faces, requiring 60 numbers rather than 108 coordinate values written separately for every triangle. Formats such as Wavefront OBJ use v records for vertices and f records for indexed faces. UV coordinates and normal vectors can supplement this core representation for texturing and smooth shading. section 11.2.1
v 0.0 0.0 0.0 v 0.0 0.0 1.0 v 0.0 1.0 0.0 v 0.0 1.0 1.0 v 1.0 0.0 0.0 v 1.0 0.0 1.0 v 1.0 1.0 0.0 v 1.0 1.0 1.0 f 1 7 5 f 1 3 7 f 1 4 3 f 1 2 4 f 3 8 7 f 3 4 8 f 5 7 8 f 5 8 6 f 1 5 6 f 1 6 2 f 2 6 8 f 2 8 4
A cube OBJ listing makes the split concrete: eight v lines define reusable points, while twelve f lines assemble them into triangles. section 11.2.1
This compact representation does not explicitly record which triangles are neighbors. That is acceptable for transformations, but operations such as mesh reduction need topological adjacency: simplifying a model means finding edges and neighboring triangles that can be changed while keeping the shape accurate enough. A half-edge representation adds references that make such neighborhood traversal explicit. section 11.2.1
Choose mesh resolution as an engineering trade-off:
- Use enough triangles to meet the accuracy required by the downstream task; a 100-micrometer printer does not benefit from a 1-micrometer model merely for its own sake. section 11.2.2
- Avoid excessive triangle counts, since larger models make processing slower and can reduce operation reliability. section 11.2.3
- Favor high-quality triangles when algorithms require them; poor-quality elements can prevent otherwise sound algorithms from working well. section 2.2.3
Distances to a triangle mesh
To build a signed-distance approximation from a triangle mesh, first measure a point’s signed distance to each triangle’s plane using a triple product for signed volume divided by a cross-product area. Use that plane distance only when the point projects inside the triangle. Otherwise measure distance to the nearest edge, and if needed to a vertex. The closest triangle supplies the distance; its plane supplies the sign. This turns a faceted boundary into an SDF that supports reliable offsets and Boolean operations. section 10.1.3 section 11.3.4
Meshes also bridge representations. Marching cubes and dual contouring convert an SDF into a triangle mesh; marching cubes generally favors smooth-surface approximation, while dual contouring generally preserves sharp features better. Conversely, meshes may be derived from parametric surfaces for 3D printing, while retaining smooth surfaces for precise machining. No representation wins everywhere: meshes trade exact smooth geometry for simpler, often faster processing. section 11.3.4 section 12.5
