Modules#

Choose a tool for your task. All seven modules have stable NumPy/CPU APIs in SplineOps 2.3, with explicit numerical contracts. See API stability and support for supported precision, dimensions, and boundaries.

Discrete samples and the continuous two-dimensional spline fitted through them.
TensorSpline

Build a continuous model from uniform samples and evaluate it at arbitrary points or grids. Reuse geometry across changing data.

Spline Interpolation
Image detail compared with and without antialiasing during resizing.
Resize

Resize N-D arrays with interpolation or projection antialiasing. Reuse fixed geometry with ResizePlan.

Resize
An image rotated using spline interpolation.
Affine

Rotate, shear, scale, and translate 2-D images or 3-D volumes. Share spline coefficients across transforms.

Affine
Gradient magnitude reveals image edges and fine structures.
Differentials

Compute cubic-spline gradients, Hessians, Laplacians, and derived features with explicit physical spacing.

Differentials
An original image, its noisy version, and the smoothed result.
Smoothing splines

Smooth real signals and arrays with fractional, radial FFT, or recursive methods. Choose the formulation and boundaries explicitly.

Smoothing Splines
A sparse spline regression fit through noisy observations.
Adaptive regression

Fit sparse piecewise-linear models to 1-D observations and inspect convergence along a regularization path.

Adaptive Regression Splines
Wavelet approximation and detail bands at multiple scales.
Multiscale

Build spline pyramids and decompose images with Haar or spline wavelets. Order-5 wavelets use documented approximate taps.

Multiscale

See the Examples for runnable illustrations and Reusable workflow recipes for workflows combining several modules.