This module explains B-splines, which form a basis for splines.
Introduces the theory and application of modern, computationally-based methods for exploring and drawing inferences from data. Covers re-sampling methods, non-parametric regression, prediction, and dimension reduction and clustering. Specific topics include Monte Carlo simulation, bootstrap cross-validation, splines, local weighted regression, CART, random forests, neural networks, support vector machines, and hierarchical clustering. De-emphasizes proofs and replaces them with extended discussion of interpretation of results and simulation and data analysis for illustration.
To perform geometric transforms on discrete images such as a rotation or zooming we need to first fit the discrete data to a continuous function. This can be done using splines. B-splines can be used to interpolate and form the continuous image from the discrete samples.
Basic introduction to splines. Link to other modules.
B-splines form a simple set of scaling functions satisfying the dilation equation with binomial filter coefficients. However B-Splines other than the zeroth order B-spline (the Haar function) are not orthogonal to its own shifts. Hence to form a perfect reconstruction(PR) filter bank system, biorthogonal scaling functions are used. Also semiorthogonal filter banks can be used to form a PR filter bank system.
