thank you for your question and sorry you didn't get an answer earlier. Most users (and developers) of scikit-image that I know of work in research labs rather than in companies. I also know a few users in national labs (such as synchrotron facilities, for example).
I would be very interested in hearing from readers of the mailing-list working outside of academia: are you using scikit-image for your work? If yes, what do you like and do not like about the package, and if not, what keeps you from using scikit-image?
A big difference with scikit-learn is that scikit-learn was the first comprehensive and successful Python package for machine learning, whereas other image processing solutions already existed for Python. In particular, openCV is an excellent package for computer vision, with good python bindings and numpy compatibility. In comparison, some of the strengths of scikit-image are its very simple and readable API (functions taking as input numpy arrays and optional parameters as keyword arguments) and its compatibility with 2D and 3D images. The documentation is also getting pretty good, with the example gallery that is a great tool for beginners (and others) to get started.
Also, machine learning is a more recent (and more rapidly growing) field than image processing, so that there is more "legacy" in the image processing field. Whereas it was easier for machine learning users to start with new technologies (Python and scikit-learn), a lot of labs have developed image processing tools in C/C++/matlab and are reluctant to switch to a new language.
But I'm sure that some companies are using scikit-image for their data processing workflow. It would be very nice to hear from them (shall we tweet about this?), in order to get feedback and maybe to set up a page like the one of scikit-learn.
Any other thoughts from the mailing-list? Cheers, Emmanuelle
On Sun, Jun 21, 2015 at 07:58:32PM +0100, 'Kevin Keraudren' via scikit-image wrote:
I just came back from PyData London, which was a great experience. I thought I would see a bit more image processing outside of an academic context, but Lyst is the only one I noticed.
I was surprised by the success of scikit-learn outside of academia , and I was wondering if people on the mailing list were aware of companies that would similarly rely on scikit-image?