sure there is, look at the docs:
Mat result; // k result ids
Mat responses; // raw responses
Mat distances; // 1/distance ~ confidence
knn->findNearest(sample, k, result, responses, distances);
hi berak am very very thankful for your answer.......k let me tell you the full problem and the procedure which am following currently in my project.......so am currently working on ocr part in which i have to process some claim forms,so in order to extract the data from fields in the form i used this KNN with opencv(character recognition) which i got it from github repository......so if i test it with some cropped images i got some output in character wise showing like it creating some bounding boxes for each character in my input image and it displays the character output .......so what i wanted is,i need the confidence score for each individual character/overall word score......how can i do that?
Asked: 2017-07-20 02:14:51 -0600
Seen: 321 times
Last updated: Jul 20 '17
