Hi, Calculation of TP,TN,FP,F​N,Acuucrac​y,precisio​n..

I have to work on face recognition on video. Actually I converted those videos in frames.In my database 10 videos are there then each videos i convert it in to 5 frames.then pca reducing is applied n then after that for classification i used the nn toolbox (feedforword). some training and testing data are there.when i apply the nn toolbox i consider 1 video means 5 frames and 2 video means 5 frames like(1,1,1,1,1,2,2,2,2,2,...) then i have to find the tp,tn,fp,fn..then how to get that one..please help me.any code related to this calculation..

Answer 1

[Image Analyst]

This is too broad a question to answer. Take your photos. I'm sure you will know who's who in the photos so you will know if the algorithm is correct or not. Then develop a face recognition algorithm (too much for us to do for you) and test it on all your images. Then construct your ROC curve.