<td>
<ul>
 <li>Version: 0.9.0.0 (release candidate 1)</li>
 <li>NuGet packages: <b>ready</b> <i>(pre-release)</i></lI>
 <li>
   <a href="https://github.com/dajuric/accord-net-extensions/raw/master/Deployment/Documentation/Help/Accord.NET%20Extensions%20Documentation.chm"> Help (offline) </a> - <i>unblock after download!</i></li>

Accord.NET Extensions is an extension framework for Accord.NET and AForge.NET . The framework is built with extensibility and portability in mind. Most provided algorithms for image processing, computer vision and statistical analysis are made as extensions. New generic image class is not tied to any specific library and video IO library, which offers unified interface for camera capture, file and directory reading, is platform abstract.

The framework is divided in libraries available through NuGet packages. The libraries can be grouped as following:

• Accord.Extensions.Math

  Provides extensions for the 2D array, graphs, contour, point transformations. Implements parallel FFT transform. Implements group matching.

  samples: cardinal spline; contour extrema; graph path search; group matching

• Accord.Extensions.Statistics

  Provides classes and extensions for the following filters: Kalman, Particle filter, JPDAF - Joint Probability Data Association Filter. Includes 2D motion models.

  samples: Kalman filtering: basic demo; object tracking - Kalman + Camshift; Particle filtering: color object tracking; model fitting

• Accord.Extensions.Imaging.GenericImage

  Implements slim generic image class and basic extensions (arithmetics). Provides multiple color spaces and conversions between them. The class can be used in non-generic way for developers who prefer AForge's UnmanagedImage style. To get compatibility for other image types install appropriate extension - NuGet package (e.g. Imaging.BitmapInterop).

  samples: interop demo (AForge, OpenCV, Bitmap, array); demo: performance, automatic color conversion, extensions...

• Accord.Extensions.Imaging.Algorithms

  Implements image-processing and computer-vision algorithms. Provides extensions for image-processing algorithms implemented in Accord.NET framework and for some image-processing algorithms implemented in AForge.NET framework.

  samples: (see Accord.Extensions.Imaging.GenericImage)

• Accord.Extensions.Imaging.Algorithms.LINE2D

  Implements template matching algorithm (~20x faster than conventional sliding window approach).

  samples: fast template matching demo

• Accord.Extensions.Vision

  Provides computer vision algorithms: Pyramidal KLT tracker, Camshift, Meanshift.

  samples: Camshift; pyramidal Lucas-Kanade tracker

• Accord.Extensions.Core

  The core of the Accord.NET Extensions Framework. Contains classes needed for other libraries such as: collections, structures, structures for parallel processing and extensions shared across libraries.

  samples: collections: sparse matrix, circular list, history, pinned array; element caching (lazy memory cache, LRU cache); method caching; parallel array operations; extensions

• Accord.Extensions.Imaging.AForgeInterop

  Provides extensions for easy interoperability between generic image and AForge UnmanagedImage.

  samples: (see Accord.Extensions.Imaging.GenericImage)

• Accord.Extensions.Imaging.BitmapInterop

  Provides extensions for interoperability with System.Drawing.Bitmap, Point, PointF, Color and drawing extensions.

  samples: (see Accord.Extensions.Imaging.GenericImage)

• Accord.Extensions.Vision.IO

  Provides unified API for IO video access: web-camera support, various video-format reading / writing, image-directory reader. All operations are stream-like and are abstracted therefore do not depend on actual video source. The library is made in platform-abstract fashion.

  samples: basic capture; capture and recording; video extraction