User guide#
Start with resize#
Regular-grid N-D resizing provides explicit coordinates and projection antialiasing.
Choose methods, geometry, axes, boundaries, and reusable plans.
Downsample an array and select spatial axes safely.
Continuous models#
TensorSpline is a stable NumPy/CPU abstraction for evaluating a
continuous tensor-product spline at arbitrary coordinates.
Construct and query continuous spline models with multiple bases and extension modes.
More spline tools#
These stable CPU modules keep their own numerical operators and support contracts. See API stability and support for the complete support matrix.
Spline-evaluated 2-D and 3-D affine transforms.
Gradients, Laplacians, and Hessian features.
Fractional smoothing and related research methods.
Sparse one-dimensional piecewise-linear models.
Spline pyramids and wavelets.