Filtering

Question 1

Spatial filtering

Answer 1

Of a neighborhood, typically done top left, by row

Question 2

Mean filtering

Answer 2

Smoothing, convolution mask with weights sum to one.

Question 3

Gauss filter

Answer 3

Emphasizes middle pixel. Reduces noise and blurs less details than mean-filtering.

Question 4

Edge enhancing

Answer 4

Laplace, sobel

Question 5

Laplace filter

Answer 5

Edge enhancing. Adds up to zero. Negative weight around, positive in middle.

Question 6

Sobel operator

Answer 6

Approximates the first derivative. Convolution mattix matched to direction.

Question 7

Frequencies responds to

Answer 7

High: fast changes, noise
Low: smoother/slower changes

Question 8

Smoothing filter in freq

Answer 8

Lowpass

Question 9

Enhance details in freq

Answer 9

Highpass/bandpass

Question 10

Frequency image

Answer 10

Low freq in middle, higher toward edges

Question 11

Erosion

Answer 11

Removes pixels of object

Question 12

Dilation

Answer 12

Adds pixels to edges of object

Question 13

Opening

Answer 13

Erosion followed by dilation, removes isolated pixels and smoothens

Question 14

Closing

Answer 14

Dilation followed by erosion, closes small holes

Question 15

Closing

Answer 15

Dilation followed by dilation, closes small holes