Revision Sesh Flashcards

1
Q

“Salt-and-pepper” noise is image noise that appears as white and black pixels in the image. Which image filter can be used to reduce/eliminate salt-and-pepper noise? Describe how the filter works.

A

the median filter is a non-linear filter that replaces the value of a pixel by the median of the value of a pixel by the median of the values in mask of the filter.
Median = “middle value” - i.e. that value m so that half of the values in the set are less than or equal to m, and the other greater than or equal to m.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What is the median filter useful for?

A

The median filter is useful for removing/reducing certain Types of random noise (impulse noise, random occurences of black and white pixels, salt-and-pepper noise).
Will lead to less blurrinng compared to averaging.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Describe spatial filtering.

A

Kernel is moved across the image. Centre of mask is aligned with current central pixels.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What is convolution in terms of spatial filtering?

A

Mask multiplied with pixel value and summed to give new value for central pixel.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What is the purpose of smoothing filters?

A

They “smoothe” the image and used for blurring and for noise reduction.
Typically used in pre-processing for:
- removal of small details from an image (e.g. prior to object extraction)
- bridging of small gaps in lines or curves
- reducing the effect of image noise

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Describe the averaging filter (mean filter).

A

Output is simply the average of the pixel contained in the neighbourhood of the filter mask.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What are gaussian filters purpose?

A

Useful for smoothing images (and better than mean filtering).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly