Python and OpenCV 3.1

First you need Jupyter and AWS ubuntu 16.04
https://www.rosehosting.com/blog/how-to-install-jupyter-on-an-ubuntu-16-04-vps/

All the powerful feature descriptors are 3rd party extensions which need special build of OpenCV

Prior to installing OpenCV, resolve some dependencies like Tesseract, PyVTK etc.,
https://medium.com/@lucas63/installing-tesseract-3-04-in-ubuntu-14-04-1dae8b748a32
https://lucacerone.net/2017/install-tesseract-3-0-5-in-ubuntu-16-04/

Then proceed with OpenCV build
http://www.pyimagesearch.com/2016/10/24/ubuntu-16-04-how-to-install-opencv/

You will run into a snag with the OpenCV build. Here is how to resolve it:
https://github.com/opencv/opencv/issues/6016

locate the file:modules/python/common.cmake and append 2 lines of code 
find_package(HDF5)

include_directories(${HDF5_INCLUDE_DIRS})


The way to access the features has changed from OpenCV 2.4 days to OpenCV 3.1
Reference: http://www.pyimagesearch.com/2015/07/16/where-did-sift-and-surf-go-in-opencv-3/#comment-431206


Some great descriptors are DCT, GLCM, FLANN, LUCID, ORB, BRIEF, SURF and SIFT, Gabor

https://www.researchgate.net/post/I_want_to_extract_Haralick_texture_features_in_openCV2
https://stackoverflow.com/questions/19556538/how-to-find-glcm-of-an-image-in-opencv-or-numpy
http://docs.opencv.org/3.0-beta/doc/py_tutorials/py_feature2d/py_orb/py_orb.html
http://docs.opencv.org/3.1.0/d5/d51/group__features2d__main.html
http://docs.opencv.org/3.0-beta/doc/py_tutorials/py_feature2d/py_matcher/py_matcher.html
http://answers.opencv.org/question/63517/how-to-successfully-implement-a-gabor-filtering/
http://docs.opencv.org/3.0-beta/modules/imgproc/doc/filtering.html
https://gist.github.com/odebeir/5237529

Finding the best features which are rotation invariant, scale invariant, resolution invariant, robust against blurring etc., is a challenge.


Of course going with a deep-learning way of classification would imply that there is no need to explicitly do feature extraction

http://www.wolfib.com/Image-Recognition-Intro-Part-1/
https://research.googleblog.com/2016/03/train-your-own-image-classifier-with.html
https://github.com/rdcolema/tensorflow-image-classification/blob/master/cnn.ipynb

Finally, remember to run Jupyter from this location so that the cv2 version is set to 3.1.0

~/.virtualenvs/cv/lib/python2.7/site-packages$