Best代写-最专业靠谱代写IT | CS | 留学生作业 | 编程代写Java | Python |C/C++ | PHP | Matlab | Assignment Project Homework代写

Matlab代写 | Project 2 CSE 402 – Biometrics and Pattern Recognition

Matlab代写 | Project 2 CSE 402 – Biometrics and Pattern Recognition


Project 2
CSE 402 – Biometrics and Pattern Recognition

1. [10 points] Consider the local orientation values in two different regions of a fingerprint image:
The orientation values are expressed in degrees. For each region, compute the Poincaré index and
determine if there is a singular point or not. If there is a singular point, indicate if it is a core (loop) or
a delta. (All computations must be shown).
2. [20 points] The ridge pattern in a local area of a finger can be approximated by a cosine wave:
w(x, y) = Acos [2πf0
(x cosθ + y sinθ)] .
Here, w(x, y) denotes the pixel intensity at location (x, y). Generate and display ridge patterns, each
of size 600 × 600, at the following orientation (θ) values: 0◦
, 45◦
, 90◦
, 135◦
. You may set A = 80 and
f0 = 0.01. For each ridge pattern, also plot the normalized magnitude of the Fast Fourier Transform
next to it.
[Hint for using FFT in Matlab:
» fft_imgv = fft2(imgv); %This command computes the FFT of image matrix imgv
» figure; imshow(fftshift(log(abs(fft_imgv) + 1)), []) %This command is needed to visualize the FFT]
3. [20 points] Based on the gradient map method discussed in class, write a program to compute the
orientation field of a fingerprint image. The orientation should be computed for each pixel location.
(So the number of rows and columns in the orientation field matrix should be the same as that of the
image). Use the Sobel Operator to compute the x and y gradient value at each pixel location. Use a
window size of 9 × 9 when computing the orientation field value associated with a pixel location (so
value of k is 4). Run your program on the set of 10 fingerprint images available here.
Use the drawOrientation.m program to display the orientation field as an overlay on the original fingerprint image. Include these overlay images in your submission.
Note: You may not be able to compute the gradient values and the orientation field values on the border
pixels. You may set those values to 0 or some other constant number in the orientation field matrix.
4. [25 points] Recall that a minutiae set, M, is a set of 3-tupled values M = {(xi
, yi
)}, i = 1, 2 . . .NM ,
where (xi
, yi
) is the location of minutiae i, θi
is its orientation, and NM is the total number of minutiae
in M. Implement the minutiae matching method discussed in class (RANSAC method) that compares
two minutiae sets M1 and M2
, and outputs the transformation parameters t x
, t y and tθ
relating M2
with M1
, along with the number of matching minutiae pairs