Image Processing 3 + 4

Question 1

Hamiltonian

Answer 1

Perpendicular to the gradient vector.

The lines are isophotes , lines of constant gray value.

Question 2

Laplacian

Answer 2

Negative if the gradient vectors are converging
Zero if the gradient vectors are parallel
Positive if the gradient vectors are diverging

Question 3

Edge detection

Answer 3

• Compute gradient magnitude image (Sobel Kernel)
• Threshold to only show edges with a certain edgeness
• Gaussian filter (reduces effect of noise)
• Laplacian

Question 4

Discrete approximation of kernel issues

Answer 4

• Truncation error (far, low ends are discarded. solution: wider kernel)
• Discretization error (over and underestimation. solution: denser sampling)

Question 5

Median Filtering

Answer 5

• Place all the grey values of a kernel around a pixel in a row, and remap it to the value in the middle of the row
• Removes noise, doesn’t shift edgess
• Min/Max filters work similarly
• These are not linear filters, but rank filters

Question 6

Nearest neighbour interpolation

Answer 6

Value of nearest pixel

Question 7

Linear interpolation

Answer 7

Added values of weighted value of w = 1 - distance to pixel (block kernel)

Question 8

2D Linear interpolation

Answer 8

Added values of pixel times opposing area (pyramid kernel)

Question 9

Anisotropic diffusion

Answer 9

Smoothing, but not equally in all directions. Is based on the heat equation.

Question 10

Grey value opening

Answer 10

ball from below, idempotent

Question 11

Grey value closing

Answer 11

ball from above, idempotent

Question 12

Dilation/Erosion

Answer 12

Uses Minkowski addition/substraction, dual operators (one on f and other on f-1 makes the results also each other’s inverse)

Question 13

Opening

Answer 13

Erosion then Dilation (details get lost, size remains) dual operators and idempotent

Question 14

Closing

Answer 14

Dilation then Erosion (structures can fuse) dual operators and idempotent

Question 15

Geodesic dilation

Answer 15

Apply dilation in iterations, but limiting growth with a control image. Eventually no more changes.

Question 16

Residue

Answer 16

Difference between the results of morphological operations (is a morphological gradient)

Question 17

White top hat filter

Answer 17

Difference between f and opening of f (used to remove background gradient when followed by tresholding)

Question 18

Ultimate Residue

Answer 18

Erode, Reconstruct, add Residue. Iterations give the ultimate residue

Question 19

Skeleton

Answer 19

Add ballz in image. Biggest ballz is on skeleton line.

Question 20

Hit-or-Miss transform

Answer 20

S1 foreground pixels have to fit, S2 foreground pixels have to be absent.

Question 21

morphological gradient

Answer 21

Difference between the erosion and dilation of an image (looks like a fucked binary laplacian with white in the middle)

Question 22

Gibs ringing

Answer 22

Artifact van een rond, binair masker toepassen op een FT. Een geblurde versie van het masker met minder harde randen verlagen dit effect.