Accord.NET provides statistical analysis, machine learning, image processing, and computer vision methods for .NET applications. The Accord.NET Framework extends the popular AForge.NET with new features, adding to a more complete environment for scientific computing in .NET.
|Tags||machine learning computer vision svm Statistics Mathematics Neural Networks crf hcrf kinect|
|Operating Systems||Windows Linux|
Origo is shutting down soon. As result, the plans to migrate to Google Code had to be antecipated a bit, and the new home is now almost complete. It was unfortunate that the just released version had been released with the old address, but a new release will be due in May with updated links. The new home is now located at http://code.google.com/p/accord/. As a side news, Accord.NET packages are now also available through NuGet ( http://nuget.org/packages?q=Accord.NET ).
Release Notes: This release adds support for previous versions of the .NET Framework, namely .NET 3.5. It also brings Cox's proportional hazard models, several fixes, and a redesigned version of the (Hidden Conditional Random) Fields namespace.
Release Notes: This release adds support for Resilient Backpropagation (RProp) learning for HCRFs, the Goldfarb-Idnani method for solving constrained QP optimization problems, and robust estimation of fundamental matrices through RANSAC, aside from many optimizations in SVM learning and evaluation. This release also includes several bugfixes, corrections, and enhancements to prior versions of the framework, and should be a great update for anyone interested in scientific computing in .NET.
Release Notes: This release introduces the Speeded-Up Robust Features (SURF) detector, Features from Accelerated Segment Test (FAST) corners detector, Limited-memory BFGS method for non-linear optimization, and threshold models for sequence rejection in hidden Markov sequence classifiers.
Release Notes: This release introduces support for independent component analysis, a new audio architecture, and a major refactoring of the hidden Markov models namespace. The new audio architecture can be used in combination with independent component analysis to perform blind source separation of audio signals. The already comprehensive set of kernels for machine learning applications has also been expanded with sparse versions of the Gaussian, Polynomial, Laplacian, Sigmoid, and Cauchy kernels.
Release Notes: Great improvements were made to the documentation. The framework now has support for Continuous density Hidden Markov Models, Gaussian Mixtures, and Non-negative Matrix Factorization.