On Sun, Feb 12, 2012 at 17:31, Emmanuelle Gouillart wrote:

Hi Ray,

this segmentation algorithm with supervised learning really seems great! I wonder whether we could have a similar algorithm in skimage. ilastik.org is down at the moment, so I can't check their license. But there is a random forest classifier in scikit-learn that we could re-use. A difficult point is the in-painting of pixels on the image; could anyone give hints about technical solutions to do this?

It's BSD. It makes heavy use of Vigra (http://hci.iwr.uni-heidelberg.de/vigra/doc/vigra/index.html) via vigranumpy, which can be a bit difficult to build (using CMAKE) in some cases (like Mac fat binary in a virtualenv). I doubt it's worth trying to reimplement a painting/labeling UI. It's not that hard to get a demonstration version working, but for anything more serious than toy problems, it will take a fair amount of time, I think. Time would be better spent working with the ilastik project on interoperability, such as scripting its pipeline (already possible, to some extent) or making it easier to call into their filtering and classification code as a python module. CellProfiler already does this in the ClassifyPixels module: https://github.com/CellProfiler/CellProfiler/blob/master/cellprofiler/module... Best, Ray