How to calculate the precision and recall for prediction deep learning

Dear all,

How to calculate the precision and recall for my prediction result deep learning?

my groundTruth(g.mat) is logical, and my prediction(tempSeg1.mat) also logical. All as attached.

anyone can help me?

Answer 1

[Yatharth]

Hey, can you tell more about your output matrix ? if you want to do element wise comparision of the 3d matrix then you can simply run 3 for loops like this and apply the formula.

  
TP=0;FP=0;TN=0;FN=0;
      for i=1:128
          for j=1:128
              for k =  1: 64
                  if(idxx(i,j,k)==1 && tempSeg1(i,j,k)==1)
                      TP=TP+1;
                  elseif(idxx(i,j,k)==0 && tempSeg1(i,j,k)==1)
                      FP=FP+1;
                  elseif(idxx(i,j,k)==0 && tempSeg1(i,j,k)==0)
                      TN=TN+1;
                  else
                      FN=FN+1;
                  end
              end 
          end
      end

Precision = TP/(TP + FP)
Recall =  TP /(TP + FN)