MaxPol Smoothing and Differentiation Package

MaxPol is a numerical differentiation toolbox for signal and image processing

What is MaxPol Package?
MaxPol is an open source code written in MATLAB delivers a comprehensive tool for numerical differentiation. The toolbox renders variety of finite impulse response (FIR) filter kernels in closed form that can be used to approximate numerical derivatives of a given discrete signals and images.

MaxPol V1.0 supports a number of utilities including:

(1) Cutoff (lowpass) design with no side-lob artifacts (for noise-robust case)
(2) Fullband finite difference formulation
(3) Derivative matrix with four design cases of boundary fomulations
(4) Arbitrary order of differentiation
(5) Arbitrary polynomial accuracy
(5) 2D Derivative Kernels with Steering moments
(6) Intuitive examples in Signal and Image processing
Which disciplinary fields can use this package?

MaxPol package provides comprehensive numerical solutions for uniform discrete differentiation that can be of interest to broad audiences in engineering and science. In the context of image processing, the package can be used in diverse imaging applications. In general if you need to compute discrete derivatives of a uniform signal or image samples in the form of gradient, Hessian, or even high order tensor, this package can be easily used with high accuracy approximations.

In the context of image processing MaxPol can be used to estimate high order feature moments applied in edge detection, curvature estimation, image enhancement, and image sharpening. MaxPol can be also used in inverse imaging problems for image restoration such as in gradient surface reconstruction, image stitching, image diffusion, variational regularization problems applied in image in-painting, enhancement/de-noising, deconvolution problems, and many more.

The package can be also utilized as a numerical solver to the PDE problems applied in fluid mechanics, acoustics, and wave equations. It can be of interest in dynamic control systems to directly estimate state variables to avoid noise robust numerical integration.

References:

[1] Hosseini, Mahdi S., and Konstantinos N. Plataniotis. "Derivative kernels: Numerics and applications." IEEE Transactions on Image Processing 26.10 (2017): 4596-4611.

Publisher link: http://ieeexplore.ieee.org/abstract/document/7944698/

[2] Hosseini, Mahdi S., and Konstantinos N. Plataniotis. "Finite Differences in Forward and Inverse Imaging Problems: MaxPol Design." SIAM Journal on Imaging Sciences 10.4 (2017): 1963-1996.

Publisher link: http://epubs.siam.org/doi/abs/10.1137/17M1118452

For more information please also refer to the MaxPol user' guide inside the package.