The source is freely available for download at Yayi's web page under the very permissive Boost license.
This release contains:
- a new library, YayiNeighborhoodProcessing, that performs (non-morphlogical) processing on neighborhoods
- morphological leveling (by means of double reconstruction)
- hit-or-miss transform for gray level images, following the definition of P.Soille
- quantiles measurements on images and regions
- circular weighted and non-weighted mean and dispersion on regions
- major bugfixes on variants as well as more python support
- support for the widely used numpy python package (import/export images into a numpy array)
- a fix for boost 1.38 backsupport
and many other improvements!
Also, a package system was deployed in this release (MacOS bundle, Debian package, Windows installer). This will allow a simplified installation of Yayi. These packages will be released and announced soon.
The code is currently tested against Boost 1.38/1.43 on MacOSX 10.5, Boost 1.40/1.43 on Ubuntu 32bits/64bits respectively, and Boost 1.38/1.43 on Windows with Visual C++Express 2008.
Your feedback will be very much appreciated!
Raffi Enficiaud