mainpage.dox

/**
@mainpage Hatching — Stripe Patterns on Surfaces

@section intro Overview

This project implements the stripe pattern pipeline from two papers by
Knoeppel, Crane, Pinkall, and Schroeder:

1. **Globally Optimal Direction Fields** (SIGGRAPH 2013) — computes a smooth
   curvature-aligned 2-direction field on a triangle mesh via a sparse
   eigenvalue problem.

2. **Stripe Patterns on Surfaces** (SIGGRAPH 2015) — given a direction field
   and target line frequency, computes globally continuous texture coordinates
   that produce stripe patterns, handling singularities via the lArg
   interpolant and barycentric subdivision.

The result is rendered with an NPR hatching shader that modulates stripe
width by shading, producing pen-and-ink style illustrations.

@section arch Architecture

The code is organized into a core library (no graphics dependencies) and
two frontends:

- **hatching_core** — the algorithm library:
  - @ref hatching::TriangleMesh — mesh data structure with topology
  - @ref hatching::MeshGeometry — discrete differential geometry operators
    (angle rescaling, parallel transport, holonomy, cotangent weights,
    Hopf differential)
  - @ref hatching::DirectionField — globally optimal direction fields
    (Algorithms 1–3 of Knoeppel 2013)
  - @ref hatching::StripePattern — stripe pattern computation
    (Algorithms 3–7 of Knoeppel 2015)

- **Desktop app** (`src/main.cpp`, `src/renderer.cpp`, `src/camera.cpp`) —
  OpenGL 4.1 / GLFW / Dear ImGui viewer with interactive parameter controls.

- **Web app** (`web/`) — Emscripten/WebAssembly build of the core library
  with a WebGL2 renderer, for browser-based viewing.

@section pipeline Pipeline

The computation proceeds in stages:

1. **Load mesh** and build edge/face topology (@ref hatching::TriangleMesh::build_topology).
2. **Compute geometry**: angle rescaling, transport angles, holonomy,
   cotangent weights, Hopf differential (@ref hatching::compute_geometry).
3. **Compute direction field**: build energy and mass matrices, solve the
   eigenvalue problem, extract curvature-aligned field
   (@ref hatching::compute_direction_field).
4. **Compute stripe pattern**: project field onto edges (Algorithm 3),
   build stripe energy matrix (Algorithm 4), solve for the stripe field
   (Algorithm 6), extract texture coordinates (Algorithm 7)
   (@ref hatching::compute_stripe_pattern).
5. **Render**: upload per-corner texture coordinates to the GPU; the
   fragment shader evaluates cos(alpha + lArg) to produce stripes, with
   width modulated by diffuse shading for NPR hatching.

@section refs References

- F. Knoeppel, K. Crane, U. Pinkall, P. Schroeder.
  *Globally Optimal Direction Fields.* ACM Trans. Graph. 32(4), 2013.
- F. Knoeppel, K. Crane, U. Pinkall, P. Schroeder.
  *Stripe Patterns on Surfaces.* ACM Trans. Graph. 34(4), 2015.
*/