Week 6

Question 1

what is the process for filtering within frequency domain?

Answer 1

1) take fourier transform of image to form the image spectrum
2) multiply image spectrum by the transformation function (e.g high or low pass filtering) (high pass is typically brighter)
3) take inverse fourier transform of output. so we are back in spatial/time domain. this is the filtered image.

Question 2

what are the two main ways of image filtering?

Answer 2

1) during frequency domain were we multiply image spectrum with a transform function (then take inverse fourier)

2) convolution of image
- uses filter kernels

Question 3

what are some post processing techniques we can use from fourier transforming

Answer 3

• image filtering (using high and low pass)
• image restoration

-

Question 4

what is Modulation transformation function

Answer 4

a number given dependant on a systems ability to reproduce the input information to exactly mirror the output.

i.e how much degradation occurs in a system

MTF = input/output

a system with high level of MTF will be able to fully reproduce image detail. (fine detail is normally lost through noise and signal loss during transmission).

Question 5

what is the Line spread function

Answer 5

a systems ability to produce sharp edges

Question 6

what happens when we take the fourier transform of a line spread function

Answer 6

we get the MTF (modulation transformation function)

Question 7

what are two main filters used for image restoration

Answer 7

inverse filtering

• applied during frequency domain (after fourier transform)
• the transform function= inverse of MTF
• will enhance image sharpness and contrast, but also noise

Weiner filtering

• Applied after output image is formed
• Has contraints that limit the use.