*DIP L5 - Image Enhancement 2

Question 1

What is the difference between point, local and global operation?

Answer 1

An image is changed, but it depends on how much of the image is changed at a time - i.e. one pixel, group of pixel or whole image

Question 2

What is the difference between spatial domain and frequency domain?

Answer 2

Spatial: XYZ… Frequency: specific place?

Question 3

What is “segmentation”?

Answer 3

Like.. cropping. This is considered a global operation

Question 4

Can point and local operators act in frequency domain?

Answer 4

NO! They are only spatial domain!!! Global is in BOTH spatial and frequency

Question 5

Histogram averaging optimises contrast

Answer 5

True

Question 6

How do you do simple mag and simple min

Answer 6

Mag: all same values, just multiple of them
Min: take average and divide, then round

Question 7

What is the difference between zooming and magnification?

Answer 7

Zooming does not change any property of the image. Magnification changes pixel value, size and matrix size

Question 8

In Simple magnification, the increases must be multiples of an integer

Answer 8

T

Question 9

In simple magnification, are the pixel sizes increasing or decreasing?

Answer 9

Neither - they remain the same. There are just more of them created

Question 10

What are other methods of mag? Why do we use these?

Answer 10

We use these because sometimes we want to magnify images that arent integer values etc… We can use nearest neighbour, bilinear interpolation and bicubic interpolation

Question 11

What is coordinate matrix? What is the purpose of this?

Answer 11

Pixel values represented at corners. For simplicity for when we do stuff like rotate with a non-orthagonal value… You literally draw lines in the middle of the boxes of the original… (lol)

Question 12

Bicubic method uses a polynomial fir across 8 neighbours

Answer 12

True

Question 13

What is the difference between bilinear and bicubic?

Answer 13

Bilinear uses 4 neighbours, bicubic uses 16!

Question 14

For coordinate matrixes, the pixel values are located where?

Answer 14

At the intersection/corners of the lines!

Question 15

Nearest neighbour: extreme values and subtle values are lost

Answer 15

False. They are still clearly represented bcos the pixel numbers are not changing

Question 16

What is the disadvantage of using nearest neighbour? What is adv?

Answer 16

Adv is that it is fast. disadv is aliasing

Question 17

Bilinear assumes a linear relationship between pixel values

Answer 17

true- think of it biLINEAR

Question 18

In Bilinear method, extreme data is lost/devalued

Answer 18

True - because you are taking an average. i.e. Research.. you got HD but you only gonna get P on exam - Averages down to alike a credit… HD is devalued

Question 19

What is the adv/disadv of bilinear method of mag?

Answer 19

Adv: more smooth, less alisaing. Disadv: extreme values are devalued

Question 20

What are your two options for non-orthagonal image rotation?

Answer 20

1. Crop the edges off (keep same size matrix)

2. MAke the matrix larger (with blank spaces)

Question 21

How do you do non-orthagonal image rotation?

Answer 21

USE THE FORMULA!!! Those rlly weird looking ones

Question 22

Image addition: NM use?

Answer 22

For noise reduction - average all the images

Question 23

Image addition: DSA use?

Answer 23

Remove anatomy to only show contrast

Question 24

RT

Answer 24

overlaying isodose lines

Question 25

What is the difference between spatial and freq domain?

Answer 25

Spatial domain tells you where the values are, freq domain just tell you how many of those values there are

Question 26

What is the main thing that Freq domains do?

Answer 26

They turn spatial relationships into freq relationships - giving them freq, amplitude and phase

Question 27

What does a band filter do?

Answer 27

Removes freqs except in a given range

Question 28

What does morph filter do?

Answer 28

Removes noise, removes ‘spikes’ in the fourier domain