Confusion Matrix and TP, FP, FN in SIFT (python implementation)

asked 2018-03-26 10:17:53 -0600

rory gravatar image

HI! I know how to calculate Precision and Recall: TP/TP+FP and TP/TP+FN; but where exactly can I extract these values from SIFT? To match KeyPoints I used both FLANN and BF. But how do I extract from those matches True Positives, False Positive and False Negative?

Maybe it is a trivial question but besides scientific article, I cannot find what are these data in a practical implementation.

Thanks for the help

edit retag flag offensive close merge delete

Comments

1

hmmm, TP, FP, FN can only be in a binary classification. (i don't think, you have that here)

berak gravatar imageberak ( 2018-03-26 10:38:09 -0600 )edit

Thanks for the answer. What I don't get is how can I evaluate my SIFT system. Do I have to use my query Image and my Image database to check for TP, FP and FN? How can I "produce" such data?

rory gravatar imagerory ( 2018-03-26 11:04:27 -0600 )edit
2

Have a look at this paper: A Performance Evaluation of Local Descriptors, Krystian Mikolajczyk and Cordelia Schmid.

Also from the same authors: Scale & Affine Invariant Interest Point Detectors.

Eduardo gravatar imageEduardo ( 2018-03-26 11:12:52 -0600 )edit
add a comment