Equations

Question 1

Linear Normalisation

Answer 1

I_n = (I - Min) [(Max_new - Min_new) / (max-min)] + min_new

Question 2

Threshold

Answer 2

Threshold = region_mean - offset

Question 3

Within (intra) class variance

Answer 3

Weight of background = (number of pixels below threshold) / total number of pixels
Mean_background = (number of pixels x intensity levels) / number of background pixels
variance_background = [(pixel intensity-mean)^2 x number of pixels with intensity] +… / number of background pixels

Repeat for foreground

Intra (within) class variance = (background weight)(background variance) + (foreground weight)(foreground variance)

Question 4

Roberts edge detector

Answer 4

sqrt[(A-D)^2 +(B-C)^2]

Question 5

Prewitt edge detector

Answer 5

sqrt[(G+H+I)-(A+B+C)]^2+[(C+F+I)-(A+D+G)]^2]

Question 6

Sobel edge detector

Answer 6

2 3x3 masks 1,2,1, 0,0,0, -1,-2,-1 and 1,0,-1, 2,0,-2, 1,0,-1
sqrt{[(A+2 * B+C)-(G+2 * H+I)]^2 +[(A+2 * D+G)-(C+2 * F+I)]^2}

Question 7

FREI & CHEN EDGE DETECTOR

Answer 7

2 3x3 masks -1,-root2,-1, 0,0,0, 1, root2, 1 and -1,0,1, -root2,0,root2, -1,0,1

sqrt[(f1)^2+(f2)^2]

Question 8

SNR

Answer 8

(variance_signal)/(variance_noise}

Question 9

PSNR

Answer 9

Peak signal to noise ratio
10log10(MAX_I^2)/(MSE)

Question 10

Euler number

Answer 10

Blobs - holes

Question 11

4-adjacency

Answer 11

1
2. 0
3

Question 12

8-adjacency

Answer 12

3. 2. 1
4. 0
5. 6. 7

Question 13

Runcode

Answer 13

(Row#_A, Col_start_A, Col_end_A : Row#_B, Col_start_B, Col_end_B :….)

Question 14

Approx perimeter

Answer 14

N_e + sqrt(2) * N_o

Question 15

White top hat

Answer 15

A - (A0K)

Question 16

NOT

Answer 16

opposites

Question 17

AND

Answer 17

if numbers are different, 0
If numbers are both 1, 1

Question 18

AND

Answer 18

If numbers are both 1, 1
Otherwise, 0

Question 18

OR

Answer 18

If both numbers are 0, 0
otherwise, 1

Question 19

XOR

Answer 19

different, 1

Question 20

closing

Answer 20

dilation then erosion

Question 21

opening

Answer 21

erosion then dilation

Question 22

Maximum suppression of canny

Answer 22

if pixel on either side is greater than examined pixel, it goes to 0