§4.6, p. 61 — the mesh metric ℓ_uv = ‖x_u − x_v‖ satisfies the triangle inequality on every face
ProvedGeomDL.Cotangent.edge_length_triangle_inequalitygeometric-deep-learningmetricp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-bookp2o-v1triangular-mesh
Let be the faces of a manifold triangular mesh whose nodes are embedded by in a Euclidean space, and let be the induced metric. For every face , every combination of its edges satisfies the triangle inequality:
This is what makes a metric on the mesh, so that a property expressed solely in terms of is called intrinsic.
Preamble
import Mathlib import Definitions.Def_GeomDL_Cotangent_Defs open EuclideanGeometry
Formal statement
namespace GeomDL.Cotangent
variable {V P : Type*} [NormedAddCommGroup V] [InnerProductSpace ℝ V] [MetricSpace P]
[NormedAddTorsor V P]
/-- §4.6, p. 61: the metric `ℓ_uv = ‖x_u − x_v‖` induced by the embedding satisfies the
triangle inequality on every face `(u, v, q) ∈ F` of a manifold mesh, for every
combination of its edges. -/
theorem edge_length_triangle_inequality {n : ℕ} (M : TriMesh n) (x : Fin n → P)
(hmesh : IsManifoldMesh M) {u v q : Fin n} (hF : (u, v, q) ∈ M.faces) :
edgeLength x u v ≤ edgeLength x u q + edgeLength x v q ∧
edgeLength x u q ≤ edgeLength x u v + edgeLength x q v ∧
edgeLength x v q ≤ edgeLength x v u + edgeLength x q u := by sorry
end GeomDL.Cotangent
Source
Bronstein et al., Geometric Deep Learning (arXiv:2104.13478v2), §4.6, p. 61, "A metric defined in this way automatically satisfies properties such as the triangle inequality"
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.