biometrics - How to calculate False Accept and Reject Rates in pattern recognition -


i working on vein pattern recognition project based on surf algorithm , euclidean distance. have completed program find maximum , minimum distance between vein features , find match when there identical image. i.e max , min distance between 2 images zero. in case, how find far , frr. 0% or missing big concept here?

even if there slight variation wouldn't match in case, guess need have threshold value compare to. have calculate max , min distance between combination of images same hand, different hands. in case, how computer far , frr. first biometrics project , helpful if directed resource me in this. thank you.

kindly me out.

usually, it's impossible find model int training set matches sample in testing set. (the situation euclidean distance 0 impossible in reality.) use many kinds of similarity measure methods, such in case, use euclidean distance.

and should have a threshold determine whether test sample , model in training set belong same person. if distance larger threshold, reject hypothesis same person , accept otherwise. have false reject , false accept cases.
in case, maybe can avoid choosing threshold every test sample compare training models. can choose nearest training model recognition result.

to calculate far , frr, should have key labels test samples can compare recognition results keys.


Comments

Popular posts from this blog

apache - Remove .php and add trailing slash in url using htaccess not loading css -

javascript - jQuery show full size image on click -