Calculating in Matlab confidence intervals and AUC in case of Identification

Question

I'm working with identification biometric system and I already calculated recognition rates with Matlab.

Moreover, I'm working under SSPP (Single Sample Per Person) context; training is performed with 200 samples representing 200 classes/labels. Similarily, testing is performed with 200 other samples representing the same 200 classes/labels. So, I have in total 400 samples, 2 for each class; one for training and the other for testing.

The training and testing images are fed into the experimental algorithm which outputs a distance matrix containing the distance between pairwise images; the similarity matrix.

Here below an example of only 10 training and 10 testing samples; there is the distance matrix between pairwise images (or dissimilarity matrix), and the confusion matrix.

Similarity matrix

Rows (1 --> 10) : 10 images of 10 different persons (persons/ classes / labels)

columns (1 --> 10) : 10 other different images each corresponding to the person with same row number.

Example : Row 3 and Column 3 are 2 images of same person (person 3)

And the actually corresponding confusion matrix is :

Confusion matrix

My questions are :

1- How to calculate confidence intervals in Matlab in case of identification

2- Can I deal with true positive rates and false positive rates in case of Identification. If it is so, how to calculate them , how to calculate AUC and plot roc curves.

Thanks in advance.