Cross-platform .NET library for digital signal processing

Every journey begins in the mind...

• color spaces and their transformations,
• real and complex algebra,
• statistical distributions,
• special math functions,
• digital response filters,
• discrete orthogonal transforms and more.

• symbolic and graphical visualization of data,
• functional, vector and matrix analysis,
• interpolation, approximation and optimization of functions,
• numerical differentiation and integration,
• solving equations,
• matrix factorization.

• Wavelet Toolbox. Provides wide functionality for the study of discrete and continuous wavelets. The toolbox also includes algorithms for discrete one-dimensional and two-dimensional wavelet transforms of real and complex signals.
• Window Toolbox. Includes a set of tools to synthesizing and orthogonalizing window functions. It implements discrete short-time Fourier and Weyl-Heisenberg transforms (Gabor analysis) for real and complex signals.
• Image Processing Toolbox. Contains efficient algorithms to processing, correcting and analyzing 32-bit images.
• Video Processing Toolbox. Includes a set of tools to video streaming and processing.

UMapx supports only

• 32 bit types - float, Complex32 and etc (compatible with System.Numerics, NAudio and other libraries),
• 32 bit image BitmapData format - Format32bppArgb (compatible with AForge.NET, Accord.NET and etc),
• 24 bit video BitmapData format - Format24bppRgb (32 bit format not recomended).

You can build UMapx from sources or install to your own project using nuget package manager.

using UMapx.Analysis;
using UMapx.Colorspace;
using UMapx.Core;
using UMapx.Decomposition;
using UMapx.Distribution;
using UMapx.Imaging;
using UMapx.Response;
using UMapx.Transform;
using UMapx.Video;
using UMapx.Visualization;
using UMapx.Wavelet;
using UMapx.Window;

• Local Laplacian filters - NET Framework desktop application for HDR imaging.
• Portrait mode effect - High quality implementation of the portrait mode effect using Neural Networks.
• FaceONNX - Face analytics library based on deep neural networks and ONNX runtime.

UMapx is based on several separate frameworks (AForge.NET, Accord.NET, Alglib and etc). Some functions have been ported from other programming languages and their toolboxes and libraries (Fortran, MATLAB, C++, Python). The purpose of this generalization was the declarative understanding of algorithms of digital signal processing and its optimization and performance improvement. UMapx is faster than AForge.NET and Accord.NET in signal processing tasks and contains a larger set of functions for matrix analysis, linear algebra and functional analysis.

MIT

A full list of references is given in a separate file.