I want to write some code to perform an iris recognition. I've already had a iris pictures which is normalized to cartesian coordinates. I've try to match theese pics using Gabor filter and Hamming distance, but without a success.
So now I'm wondering if in openCV exists a matcher which will be good to provide this kind of recognition? I know there is some predefined matchers, but which one will be the best for very similar pictures? Let me also add that my application should decide if given iris exist already in database or not.
I'm using Java version of openCV
Feature extraction & matching to the help!
The problem is a typical recognition application. Inder to match two image successfully, you have to find some distinct or unique properties that will help you to find what you are looking for. These are called "Features" in pattern recognition/image processing environments. However, that is not simple as it is. There are many feature extraction methods which are already implemented in OpenCV (SIFT, SURF, ORB etc.)
In Iris recognition case, you will need scale, illuminance and rotation invarient features. SIFT features will be the best candidate for this work. Therefore, extract SIFT features of iris images and store them to the database. OpenCV also has feature matchers, flaNN is one of them.
OpeCV
SIFT: http://docs.opencv.org/3.1.0/da/df5/tutorial_py_sift_intro.html#gsc.tab=0
FLANN: http://docs.opencv.org/2.4/doc/tutorials/features2d/feature_flann_matcher/feature_flann_matcher.html