CS 659辅导、讲解Matlab、辅导CS 659 Image Processing

2019.02.05 - 首页 >> Matlab编程

Dr. Frank Shih

CS 659 Image Processing

Homework #1

Covering Lectures 1, 2, 3 and Familiarize with Matlab Image Processing Toolkit

NOTES: Submit only the homework “solution” (do not include these homework questions) in

Microsoft Word format to http://njit2.mrooms.net/ before the above deadline. Absolutely, no late

submission is accepted. Write the answers in your own words individually. Any plagiarism will

post a “ZERO” score or cause a “FAIL” grade. The course requires Matlab programming. You can

download the Matlab software from the NJIT website: http://ist.njit.edu/software/download.php. If

you download it off campus, you will need to install VPN license software before downloading (See

ist.njit.edu for download instruction). Remember to choose the newest student version with your

own OS, and after starting MATLAB, request Product Key to activate it.

Notes:

1. In MATLAB window, choose “Preferences”. Click “Image Processing” selection. In the “Border

Style”, choose “Tight” instead of “Loose”. Click “OK”.

2. When the resulting image is shown in another window, click “Edit” and choose “Copy Figure”.

Then go to word file, paste the image into your file.

3. Submit your answer in a Microsoft Word file. Do not include the questions; just provide the

required answers in the file.

4. Totally, there are 5 questions. Each question is 20 points. Grading policy checks the correctness

and completion of showing resulting images, Matlab codes, and text responses.

5. Required images: barbara.jpg, Lena.jpg, cameraman.jpg, lena_256.bmp

Check List:

1. Listen my lecture videos on Moodle: Lectures 1, 2, 3.

2. Download Matlab in your PC and learn how to write its code for processing images.

3. Use internet resources to learn details. Read “barbara.jpg” by imread and show it by imshow.

4. Learn how to show image size and how to convert a color image into grayscale by rgb2gray.

5. Learn to use for-loop to cut out the central rectangle 300x300 of the image and the other

remaining pixels are set to be 32 using cut.m.

Script file: cut.m

fa=imread('barbara.jpg');

for i=1:576

for j=1:720

fc(i,j)=32;

end

for i=139:438

for j=211:510

Dr. Frank Shih

fc(i,j)=fa(i,j);

end

imshow(fc/255)

6. Read a youtube video on Matlab Image Processing Basic at:

https://www.youtube.com/watch?v=3St0byWYcEI

1.1 (20 points)

(a) Use a cell phone or digital camera to take your face photo and display it.

(b) Use Matlab to crop the above photo into a small image showing only your face skin portion.

(d) Show the histogram of the above cropped grayscale image.

1.2 (20 points)

(a) Change the grayscale image 'barbara_gray.jpg' into a binary image, which has only two values: 0

and 1. Let the gray level greater than 50 be 1 (white) and otherwise be 0 (black). Give the

Matlab code and show the binary image.

(b) Select the gray level which is greater than 200 or smaller than 100 to be 1 (white) and otherwise

be 0 (black). Give the Matlab code and show the binary image.

Otsu's method: graythresh, and use the function im2bw to change the grayscale image into a

binary image by the Otsu’s threshold value. Show the binary image. What is the Otsu’s

threshold value between 0 and 1? What is the equivalent Otsu’s threshold value between 0 and

255?

1.3 (20 points)

(a) Read the 512 x 512 Lena image as F and crop its lower-right quadrant (row from 257 to 512 and

column from 257 to 512) into a 256 x 256 “Lena_256” image as P. Write P into a new image file

by imwrite(P, 'Lena_256.jpg', 'jpeg'). Show the cropped image 'Lena_256.jpg'.

(b) Read the cropped Lena_256 image as P and add 80 to each pixel and show the image using

imshow(P+80)

previous one?

(d) Multiply P by 2. R=P*2; figure, imshow(R).

(e) Divide P by 3. S=P/3; figure, imshow(S).

(f) Read the 256 x 256 “cameraman.jpg” into I and add both “Lena_256” and “cameraman.jpg”

together by J=imadd(P, I). Show the combined image.

Dr. Frank Shih

1.4 (20 points)

(a) Read the following image: lena_256.bmp which is 256 ? 256 resolution. Use program to

decrease the resolution to 128 ? 128 and 64 ? 64. Show the above three images. An example of

Matlab code: x=imread('lena_256.bmp'); x2=imresize(x,1/2); x3=imresize(x2, 1/2);

Provide your source code and show the three images.

(b) Display all above three images as 256 ? 256. In other words, the image of 128 128 will be

resized with the parameter 2 rather than 1/2. All the pixels are repeated to produce an image

with the same size as the original. In Matlab, we can use: m2=imresize(imresize(x,1/2),2);

m3=imresize(imresize(x2,1/2),4); Provide your source code and show the three images.

and show that it

performs averaging operation which results in blurring of the image. Also analyze the impact of

increasing the size of the mask to

; that is, the mask is. Provide

your source code and show the three images. Note: use MATLAB functions “ones” to generate

the matrix and use “conv2” to do the convolution.

1.5 (20 points)

(a) Refer to Lecture Module 1, Topic: Image Transforms, Powerpoint Slide, pages 15&16. Hand

calculate to perform the linear convolution between these two matrices x(m,n) and h(m,n) given

below. Let the outside boundary pixels be zeros. You need to expand the boundary of x(m,n) to

obtain all non-zero values in the result. Let the upper-left pixel of x(m,n) and h(m,n) be the origin.

You need show step-by-step hand calculation process.

(b) Similarly to (a), but hand calculate to perform the linear correlation.

(d) Convolution in spatial domain is equal to multiplication in the frequency domain. Prove this by

a MATLAB code. Show the code and resulting images.

1 2 3

( , ) 4 5 6 ( , ) 3 4 5

7 8 9

x m n h m n

Survey Questions:

(a) Describe your previous BS/MS major and educational background.

(b) Describe your technical programming skills and languages used. Do you know Matlab

Dr. Frank Shih

programming? If not, you can start learning it by yourself from related books.

image processing applications if you had before.