Assig

CS631T – Spring 2019
Assignment 3 – Image Enhancement in the Spatial Domain

Due Friday, 02/22/2018
Submit electronically before 11:59pm on Blackboard
Total Points: 50 points
Part 1: Matlab Coding [20 points]

Save all solutions in a single m-file. Be sure to place semicolons wherever appropriate, to
suppress unnecessary console output, such as when loading images into memory, or operating on
them. Please submit only your m-file via the BB system. Do not send any image!
Please include comments at the top of each m-file. The comments should contain your name,
your email and assignment number. In your main function, place a message “-----Finish
Solving Problem X-----” followed by a pause command (i.e., wait for a key to be pressed
before continuing) at the end of each solution, where X is the question number (i.e., 1, 2, or
3).
Problem1: Exercises on Low-pass and High-pass Filters in the Spatial Domain [20 points]
(If boundary extension is needed, please pad the boundary with 0’s. Convert the image
between unit8 and double as needed)
a) Implement a MeanFilter function to perform a filtering operation on the input image.

[Note: the input of the MeanFilter function should be an image array and a mask. The
output should be an image array.]
Call this function to process the noisy image Circuit by using an average 3-by-3
averaging filter and a standard 5-by-5 averaging filter, respectively. You are not allowed
to simply use the Matlab “filter2” or “conv2” function in your function implementation.
Display original image and two processed images in figure 1 with the appropriate titles.
[3 points]
b) Use Matlab function fspecial and filter2 to perform the same mean filtering operation on
the Circuit image. Process the image with a 3-by-3 averaging filter and a 5-by-5
averaging filter, respectively. Compare the two images with the results from your own
MeanFilter function. Display if they are same. [2 points]
c) Implement a MedianFilter function to perform a filtering operation on the input image.
[Note: Both input and output of the MedianFilter function should be an array with data
type uint8.]
Call this function to process the same noisy image Circuit by using a weighted 3-by-3
median filter M = [1 2 2; 1 1 1; 2 1 1] (see definition below) and a standard 5-by-5
median filter, respectively. You are not allowed to simply use the Matlab “medfilt2”
function in your function implementation. Display original image and two processed
images in figure 2 with the appropriate titles. [5 points]
The standard median filter explained in the class is a special kind of the weighted median
filter. Each value in M indicates the number of copies of the corresponding masked
value in the image involved in the standard median filtering.
d) Use the Matlab function medfilt2 to apply a median filter on the Circuit image. Process
the image with a 3-by-3 median filter and a 5-by-5 median filter, respectively. Compare
with the resultant images from your own MedianFilter function and display if they are
same. [2 points]
e) Use the Laplacian mask to filter the image Moon by calling an appropriate Matlab
function. Find the function by yourself. Use the formula Enhanced Image = Original
Image – Filtered Image to get the final enhanced image. Use imshow to display four
images including the original image, the filtered image, the scaled filtered image, and the
enhanced image, in a new figure with the appropriate titles (Refer to the Figure 3.38 in
the textbook). [3 points]
f) Close all figures and all variables in the workspace.
Part2: Exercise Problems. Submit the answers of this part as a word file or a pdf file
separately from the Matlab m-file above. [30 points]
Problem1: Suppose we apply the following operations to input image f:

a. 1x3 mean filter (centered about origin) applied to f to give output f1.
b. 3x1 mean filter (centered about origin) applied to f to give output f2.
c. Take the mean (average) of f1 and f2 to get the final output.
These three operations can be done all at once with one mask operator. Derive this operator,
carefully specifying the origin and mask weights. You must justify your answer.
Problem2: Given original image:
And mask
a) What is the resultant image of convolution (as defined in Eq. (3.4-2))?

b) What is the resultant image of correlation (as defined in Eq. (3.4-1))?
Problem3:
a) Suppose that you filter an image, f(x, y), with a spatial filter mask, w(x,y), using
convolution, as defined in Eq. (3.4-2), where the mask is smaller than the image. Prove
the important property that, if the coefficients of the mask sum to zero, then the sum of
all the elements in the resulting convolution array will be zero also. You may assume that
the border of the image has been padded with multiple layers of zeros.
b) Would the result to (a) be the same if the filtering is implemented using correlation, as
defined in Eq. (3.4-1)?

Assig

Uploaded by

Copyright:

Available Formats

Assig

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Assig

Uploaded by

Copyright:

Available Formats

CS631T – Spring 2019

Assignment 3 – Image Enhancement in the Spatial Domain

Part 1: Matlab Coding [20 points]

a) Implement a MeanFilter function to perform a filtering operation on the input image.

Problem1: Suppose we apply the following operations to input image f:

Problem2: Given original image:

a) What is the resultant image of convolution (as defined in Eq. (3.4-2))?

You might also like