General Information:

•    This is a face-to-face closed book examination.

•    Scope: Chapters 5 to 7.3

•    1 hour in-class exam: Nov. 7, 2022. 8:40pm – 9:50pm.

Topics

1.   Image Restoration and Reconstruction

(a) Know noise removal filters: Arithmetic mean filter, geometric mean filter, median

filter, alpha-trimmed mean filter, adaptive median filter

(b) Know three ways for estimating impulse response: by observation, experimentation

and modelling.

(c) Know inverse filtering

•    What issue does it have?

•    How to mitigate this issue?

(d) Know Wiener filtering

•    What does it optimize?

•    What issue does it have?

•    How to mitigate this issue?

(e) Image Reconstruction from Projections

•    Know Radon Transform

•    Define “sinogram” . Why is this representation called sinogram?

•    Describe backprojection. Describe mathematically why backprojection results in a blurred image.

•    Know the Fourier-slice theorem

•    Define “filtered backprojection” . Describe the filter that should be used mathematically. List the steps involved in filtered backprojection.

2.   Geometric operations

(a) Linear transformation: expected to know formulae to relate coordinates in the input

image with those in the output image

•    Translation

•    Scaling about origin or an arbitrary point

•    Rotation about origin or an arbitrary point

•    Composite transformation. Order of operation does matter.

(b) For all transformation described in (a), know how to define the affine matrix to

implement the transformation using Python built-in tool [i.e., need to be able to determine T passed to cv2 .warpAffine(im, T, (width, height)) for the  transformation described in (a).]

(c) Define forward mapping and backward mappings. Why are the advantages of using backward mapping? Need to be able to perform the backward mapping given the    forward mapping [i.e., need to be able to express (i,j) in terms of (i’,j’).]

(d) Grey level interpolation

•    Nearest neighbour

•    Bilinear