Reference no: EM132336791

Biometrics

Question 1. Open the file, s048r.txt, provided with this assignment (tab delimited). This file contains a verification test result. It contains two columns, test.subject and test.out. test.subject is the true label, and test.out is the prediction result. You may use any tools or language you would like including excel. Answer the following questions based on the prediction result in this file. *Note that the positive class here is s048.

a. Construct a confusion matrix.

b. What is the accuracy of this model? Is this a useful measure to evaluate the model?

c. Compute FMR, FNMR, Precision, and Recall.

Question 2. Answer the following questions in your own words.

a. How are singularities used in fingerprint recognition?

b. What is the thinning process in fingerprint feature extraction? And what benefit do they have?

c. Why do we need to find local ridge orientation and frequency earlier on in the processing of fingerprint image?

Question 3. Perform a singularity detection in the following data. Use the definitions used in week 3 live session slides. Your answer should include all missing values in the table and the type of singularity detected.

Use the following two equations to fill in the column titled δ and the column titled Δ.

δ (k)= θ ((k +1)mod N )- θ (k)

            δ (k)     if |δ (k)|< Π/2
Δ(k) =   δ (k)+Π  if δ (k) ≤ -Π/2
           δ (k)-Π   if δ (k) ≥ Π/2

Once you have filled in the table give the type of singularity you believe is being represented based on:

                         360, then whorl

                         180, then loop

∑_k∈(0...7} Δ(k) = - 180, then delta

                          0, then singularity

Question 4. Determine the 3x3 binary pixel grid for:

a) A bifurcation point;

b) A non-minutiae point
(In live session we displayed and discussed the grid for the termination case).

For each specify:
- The values of b0 , ... , b7 for each case.
- What are their crossing numbers?

Crossing Number = ∑_i∈(1...7)|b_i - b_(i+1)mod8|

Question 5. The following image shows the values in grayscale. Perform the necessary steps to detect minutiae points. You don't need to detect any minutiae centered at the edge. Show your steps. Your result will include the coordinate of detected minutiae points and their types.

Note: 0 represents the darkest shade, higher numbers represent brighter shades.

Attachment:- s048r 1.rar