Week 1 - Image Processing

Question 1

What are the four ways we can represent an image

Answer 1

Image Function
Landscape
Array of pixels
Image histogram

Question 2

What is a grayscale image function

Answer 2

I : R^2 - > R
(x,y) -> I(x,y)
Gives the Intensity at position I(x,y)

A digital image is a discrete (sampled) version of this function

Question 3

What is an image landscape

Answer 3

Imagine throwing a piece of cloth over a scene
creates a smooth and continuous landscape

Question 4

What is an image histogram

Answer 4

Count the amount of pixels at each grey level

Question 5

What is Spatial Resolution

Answer 5

The number of pixels that determines the resolution
(want more detail -> need more pixels)
eg Pixels per inch
-Allows for smaller scale structured images (to be seen in detail)
-constantly increases in digital cameras

Question 6

What is grey level resolution

Answer 6

Number of different shades of grey that can be represented in an image

Higher resolution - > captures fine changes in brightness

Question 7

What is important about having an array of pixels representation

Answer 7

We get the spatial relationship

Question 8

What is Image arithmetic

Answer 8

Considering images as functions
g(x,y) = f(x,y) + 20
brightens by 20

g(x,y) = f(-x,y)
image is inverted along the y axis

Question 9

What is image addition and what does it do visually

Answer 9

Takes the average over images in a sequence
(adding the two corresponding pixels)
Reduces noise

Question 10

What is image subtraction and what does it do visually

Answer 10

Takes the difference of two images
Can capture a movement between two static background (the leftover is the moving object in the image)
Captures, shadows, reflections

Question 11

What is Noise

Answer 11

A stochastic (random) process
All images contain noise

Question 12

Do identical images have the same noise

Answer 12

No

Question 13

How do we model noise

Answer 13

As small fluctuations
Noise is normally distributed
Some small fluctuations may actually be small changes in an image and not noise

Question 14

What is the signal-to-noise ratio

Answer 14

SNR = max signal / σ

σ has subscript noise
Larger σ creates a smaller ratio and a more noisy image

Question 15

How do we reduce noise by temporal averaging

Answer 15

We can average noise over N images

σ(subscript N) = σ / √N

Take N consecutive samples of the same image
the noise reduces

Question 16

How do we reduce noise by spatial averaging

Answer 16

(When we do not have the choice of sampling an image n times)
Pass a 3x3 mask over an image, replacing the central pixel with the average of neighbours
“local averaging”

Question 17

What is the issue with reducing noise by spatial averaging

Answer 17

Lose resolution - like blurring 5x5
particularly along edges

Question 18

What is sub-sampling

Answer 18

Throw away every other row and columns in the image to create a 1/2 size image
Creates a very scruffy output

Question 19

How can we improve sub-sampling

Answer 19

Local average first (which removes small scale detail and noise) then subsample
Means we are not keeping noise

Question 20

What are the 3 ways we can define noise using its neighbours

Answer 20

Noisy dark areas
-just a messy area of dark pixels

Just noise
-white pixels with one dark pixels (one noise pixel)

Vertical edge
-noise is aligned vertically

Question 21

What is neighbourhood processing and the 2 main methods

Answer 21

Modifying the pixel values of an image based on the values of neighboring pixels in a window

1) Rank filtering
2) Convolution

Question 22

What is Rank filtering

Answer 22

Takes values in the filter window and orders them darkest to lightest
There are different ways to rank: minimum (erosion), maximum (dilation), median filtering
Replace the middle pixel with whichever chosen method (minimum/median/maximum)
Move to the next window and repeat

Question 23

What is the effect of median filtering

Answer 23

Removes both dark and light noise spots
It is not the same as taking the mean
Mean -> creates shades of grey
Median - > produces only either white or black, keeps sharp edges
Works very well for salt and pepper noise

Question 24

What is Convolution

Answer 24

Very often used in pre-processing
Want to maintain the spatial patterns
Uses a weighted filter window

Eg 1 2 1

applied to input image: 0.5 1 1

only changing the middle value to the average of all
The middle value = 0.9

NOTE divide by 4 (total of filter weight) not by pixel number

Question 25

What is a kernel

Answer 25

filter/mask/convolution mask/ window

Question 26

What is asterisk notation for convolution

Answer 26

I~ = g * I
or
I~ = I * g

g = kernel weights
I = input image
I~ = output image

Question 27

Why does convolution create a cropped output image

Answer 27

We cannot begin in the corner pixel as it does not have a full set of 3x3 neighbors

Question 28

What are 3 padding methods

Answer 28

Zero: set all pixels outside image to 0

constant (border colour): set all pixels outside source image to specific value

mirror: reflect pixels across the image edge

Question 29

What is linear filtering

Answer 29

output is a linear combination of the input pixel values within a local neighborhood

Question 30

What is a smoothing kernel

Answer 30

9x9 mask filled with 1s
(constant mask)
___|-|___

Question 31

What is a gaussian kernel

Answer 31

Smoothing using a normal distribution
Any kernel representing a single bump distribution will blur the image
1 2 1
2 4 2
1 2 1

Question 32

How does changing sigma affect the gaussian filter

Answer 32

small σ = sharper, more weight to central pixels, less blur

large σ = weight spread, more blurred

Question 33

What is segmentation

Answer 33

The division into background pixels and object pixels
grayscale image -> binary label image
Only applicable in simple, high contrasting images

Question 34

What is thresholding

Answer 34

b(x,y) = (g(x,y) < T)
If we want the dark pixels (objects of interest)
b(x,y) is a binary image
We can use two thresholds if we want two different gray levels

Question 35

What is a bimodal image histogram

Answer 35

A histogram that shows two distinct levels
background/foreground

Question 36

What is the intermodal minimum

Answer 36

Where we place the threshold
marks in between background and foreground pixel values

Question 37

Histogram as a derivative of area

Answer 37

H(D) = dA(D) /dD

where D = grey level
Because the area under the histogram curve represents the total intensity

Question 38

What is Adaptive Thresholding

Answer 38

When we encounter shaded backgrounds and a single threshold wont work
Smooths first to get estimate of varying background
Subtract smoothed image from the original
Then threshold

Question 39

What is over/under segmentation

Answer 39

Thresholding too high(over) or low(under) causing a poor quality output image

Question 40

How do we carry out automated thresholding selection

Answer 40

There are many algorithms that do this

Question 41

What does having longer words at each pixel allow

Answer 41

Each pixel can display a greater range of grey values
(bits per pixel)

Question 42

what is a word

Answer 42

basic unit of data that can process a single operation
size of word = __bits
commonly 1 byte = 8 bits

Question 43

What do lighter pixels in grayscale images represent

Answer 43

objects (why dilation = choosing maximum Intensity pixel)

Question 44

Is convolution linear filtering

Answer 44

Yes

Question 45

Is median filtering linear filtering

Answer 45

No