Class: LaplacianSmoothing

Defined in: utils/LaplacianSmoothing.ts:28

Laplacian smoothing utility for mesh geometry

Provides algorithms to smooth mesh surfaces by moving vertices toward the average position of their neighbors. Useful for:

• Reducing noise in scanned surfaces
• Smoothing procedurally generated meshes
• Creating organic, flowing surfaces

LaplacianSmoothing

Example

// Standard Laplacian smoothing
LaplacianSmoothing.smooth(vertices, faces, 3, 0.5);

// Volume-preserving Taubin smoothing
LaplacianSmoothing.taubinSmooth(vertices, faces, 3, 0.5, -0.53);

// Apply to THREE.js geometry
LaplacianSmoothing.smoothGeometry(geometry, 3, 0.5, 'taubin');

Constructors

Constructor

new LaplacianSmoothing(): LaplacianSmoothing;

Returns

LaplacianSmoothing

Methods

smooth()

static smooth(
   vertices, 
   faces, 
   iterations?, 
   lambda?, 
   boundarySmoothing?
): Float32Array;

Defined in: utils/LaplacianSmoothing.ts:72

Apply standard Laplacian smoothing to vertices

Algorithm: Each vertex is moved toward the average position of its neighbors. This tends to shrink the surface slightly with each iteration.

Parameters

vertices

Float32Array

Vertex positions (will be modified in place)

faces

Uint32Array

Face indices defining mesh topology

iterations?

number = 1

Number of smoothing passes (1-10 typical). More = smoother but may lose detail

lambda?

number = 0.5

Smoothing strength per iteration:

• 0: No smoothing
• 0.1-0.3: Gentle smoothing, preserves features
• 0.4-0.6: Moderate smoothing
• 0.7-1.0: Aggressive smoothing, may over-smooth

boundarySmoothing?

boolean = false

If false, boundary edges remain fixed (recommended for open meshes)

Returns

Float32Array

The smoothed vertices (same array, modified in place)

Example

// Gentle noise reduction
LaplacianSmoothing.smooth(vertices, faces, 2, 0.25, false);

// Aggressive smoothing
LaplacianSmoothing.smooth(vertices, faces, 5, 0.7, false);

---

taubinSmooth()

static taubinSmooth(
   vertices, 
   faces, 
   iterations?, 
   lambda?, 
   mu?, 
   boundarySmoothing?
): Float32Array;

Defined in: utils/LaplacianSmoothing.ts:193

Apply Taubin smoothing (volume-preserving algorithm)

Algorithm: Alternates between shrinking (positive lambda) and expanding (negative mu) passes. This prevents the volume loss common in standard Laplacian smoothing.

Parameters

vertices

Float32Array

Vertex positions (will be modified in place)

faces

Uint32Array

Face indices defining mesh topology

iterations?

number = 1

Number of shrink-expand cycles (1-5 typical)

lambda?

number = 0.5

Shrinking factor (0.3-0.7 typical). Higher = more aggressive

mu?

number = -0.53

Expansion factor (-0.2 to -0.7 typical). Must be negative.

• Classic Taubin: mu = -(lambda + 0.02)
• Common default: mu = -0.53 when lambda = 0.5

boundarySmoothing?

boolean = false

If false, boundary edges remain fixed

Returns

Float32Array

The smoothed vertices (same array, modified in place)

Example

// Classic Taubin parameters
LaplacianSmoothing.taubinSmooth(vertices, faces, 3, 0.5, -0.53, false);

// Gentle volume-preserving smoothing
LaplacianSmoothing.taubinSmooth(vertices, faces, 2, 0.33, -0.35, false);

---

smoothGeometry()

static smoothGeometry(
   geometry, 
   iterations?, 
   lambda?, 
   method?, 
   boundarySmoothing?, 
   mu?
): void;

Defined in: utils/LaplacianSmoothing.ts:218

Apply smoothing to a BufferGeometry

Parameters

geometry

BufferGeometry

iterations?

number = 1

lambda?

number = 0.5

method?

"laplacian" | "taubin"

boundarySmoothing?

boolean = false

mu?

number = -0.53

Returns

void