10/27 Notes

Minutiae Matching

P = {(x_i, y_i, theta_i)} i = 1 -> N

Q = {(x_j, y_j, theta_j)} j = 1 -> M

T:P->Q congruent T(P)->Q

(T is the transformation operation)

T_theta = the rotation

T_x = x translation

T_y = y translation

See Fig 1 in magna Notebook

RANSAC Algorithm

Stands for Random Sample Consensus

See Fig 1 in magna Notebook

This Transforms all points in P according to t_x, t_y, t_theta, x_c, y_c: so P -> P’

Then we check the number of correspondences between P’ and Q (he wants us to think about how to do this)

There will obviously be a tolerance for how close a point can be

Alg

N_max = the number of points in correspondences = 0

for i = 1 to N

for j = 1 to M

-t_x = x_j - x_i

t_y = y_j - y_i

t_theta = theta_j - theta_i

for k = 1 to N

See fig 1 in magna Notebook

end for loop

compute number of corresponding points between P’ and Q - N_y

if N_y > N_max

N_max = N_y

end for loop

end for loop

compute score so S = (N_max)^2 / (NM) x 100