Image Processing using Matlab

Sumitha Balasuriya
http//www.dcs.gla.ac.uk/~sumitha/
 Images in Matlab
• Matlab is optimised for operating on
matrices
• Images are matrices!
• Many useful built-in functions in the
Matlab Image Processing Toolbox
• Very easy to write your own image
processing functions

2
 Loading and displaying images
>> I=imread('mandrill.bmp','bmp'); % load image

image format
Matrix with image filename as a string
image data as a string
>> image(I) % display image
>> whos I

Name Size Bytes Class

I 512x512x3 786432 uint8 array

Grand total is 786432 elements using 786432 bytes Matlab can only perform
Display the left
arithmetic operations on half of the
Dimensions of I (red, green
data with class double! mandrill image
and blue intensity information)

3
 Representation of Images
• Images are just an array of numbers
>> I % ctrl+c to halt output!

• Intensity of each pixel is represented by the pixel element’s

value in the red, green and blue matrices
>> I(1,1,:) % RGB values of element (1,1)
ans(:,:,1) =
Red
135 Images where the pixel value in the image
ans(:,:,2) = represents the intensity of the pixel are called
intensity images.
97 Green

ans(:,:,3) =
33 Blue

4
 Indexed images
• An indexed image is where the pixel values are indices to elements in a colour map or
colour lookup table.
• The colour map will contain entries corresponding to red, green and blue intensities for
each index in the image.
>> jet(20) % Generate a jet colourmap for 20 indices 3 4 7 3 6 19 8 9 1 2
ans =
5 6 14 4 2 5 6 1 4 5
0 0 0.6000
0 0 0.8000
RGB Entry for index value 3
2 8 9 4 2 13 7 8 4 5
0 0 1.0000
0 0.2000 1.0000 5 1 11 5 6 4 1 7 4 4
0 0.4000 1.0000
1 9 5 6 5 5 14 4 6 5
0 0.6000 1.0000
0 0.8000 1.0000 5 9 2 1 11 1 3 6 1 9
0 1.0000 1.0000
0.2000 1.0000 0.8000 7 6 8 18 1 8 1 9 13 3
0.4000 1.0000 0.6000 9 2 3 7 2 9 8 16 6 4
0.6000 1.0000 0.4000
0.8000 1.0000 0.2000 7 8 6 7 4 15 8 2 1 3
1.0000 1.0000 0 Values can range
1.0000 0.8000 0
7 5 10 8 4 10 4 3 6 4
from 0.0 to 1.0
1.0000 0.6000 0
1.0000 0.4000 0
1.0000 0.2000 0
Red, green and blue intensities of
1.0000 0 0 the nearest index in the colourmap
0.8000 0 0
0.6000 0 0
are used to display the image.
Image Processing using Matlab Sumitha Balasuriya 5
 Displaying indexed images
>> I2=I(:,:,2); % green values of I
>> image(I2) Matlab considers I2 as an indexed image as it doesn’t
contain entries for red, green and blue entries

>> colorbar % display colourmap

Index

Associated
color

Colour
Lookup
Table

6
 Displaying indexed images (continued)

Red =1.0,
Green = 1.0,
Blue =1.0,
corresponds to
• change colourmap index 64

>> colormap(gray)
Red =0.0,
Green = 0.0,
Type >>help graph3d to get a list of built-in Blue = 0.0,
colourmaps. Experiment with different corresponds to
built-in colourmaps. index 1

Define your own colourmap mymap by Red =1.0,

creating a matrix (size m x 3 ) with red, Green = 1.0,
green, blue entries. Display an image using Blue =1.0,
your colourmap. corresponds to
index 255

• scale colourmap Red =0.0,

Green = 0.0,
>> imagesc(I2) Blue = 0.0,
corresponds to
index 0

7
 Useful functions for displaying images
>> axis image % plot fits to data
>> h=axes('position', [0 0 0.5 0.5]);
>> axes(h);
>> imagesc(I2)

Investigate axis and axes

functions using Matlab’s help

8
 Histograms
• Frequency of the intensity values of the image
• Quantise frequency into intervals (called bins)
• (Un-normalised) probability density function of
image intensities

9
 Computing histograms of images in Matlab

>>hist(reshape(double(Lena(:,:,2)),[512*512 1]),50)

Histogram Convert image into a 262144 by Number of bins

function 1 distribution of values

Generate the histograms of the green channel of the Lena image using the
following number of bins : 10, 20, 50, 100, 200, 500, 1000

Histogram equalisation works by equitably distributing the pixels among the

histogram bins. Histogram equalise the green channel of the Lena image
using Matlab’s histeq function. Compare the equalised image with the
original. Display the histogram of the equalised image. The number of pixels
in each bin should be approximately equal.

10
 Visualising the intensity surface
>>surf(double(imresize(Lena(:,:,2),[50 50])))

Change type to Remember to reduce

double precision size of image!

Use Matlab’s built-in mesh and

shading surface visualisation
functions

11
Useful functions for manipulating
images
• Convert image to grayscale
>>Igray=rgb2gray(I);
• Resize image
>>Ismall=imresize(I,[100 100], 'bilinear');
• Rotate image
>>I90=imrotate(I,90);

12
 Other useful functions
Convert polar coordinates to Check if a variable is null Trigonometric functions
cartesian coordinates >>isempty(I) sin, cos, tan
>>pol2cart(rho,theta)
Convert polar coordinates to Find indices and elements in a Fast Fourier Transform
cartesian coordinates matrix
fft2(I)
>>cart2pol(x,y) >>[X,Y]=find(I>100)

Get size of matrix Change the dimensions of a Discrete Cosine Transform

>>size(I) matrix dct(I)
>>reshape(rand(10,10),[100 1])
Add elements of a Matrix Exponentials and Logarithms
(columnwise addition in matrices) exp
>>sum(I) log
log10

13
 Convolution
Bit of theory! Convolution of two functions f(x) and g(x)

h( x )  f ( x )  g ( x )  

f (r ) g ( x  r )dr

Image convolution Filter Support region

Output of filter where
filtered image operator (mask/kernel)
g(x-r) is nonzero

Discrete image processing 2D form

height width
H ( x, y )    I (i, j )M ( x  i, y  j )
j 1 i 1
Compute the convolution where
there are valid indices in the kernel

Image Processing using Matlab Sumitha Balasuriya 14

 Convolution example
Image (I)
0 0 0 0 0 0 0 0 0 0
197 199 195 194 189 190 132 90 112 101 Filter (M) 0 196 196 194 192 170 137 105 97 0
194 194 198 201 189 196 150 85 87 97
0 195 196 194 192 167 133 98 92 0
194 194 198 195 186 191 109 90 90 124
0 194 194 193 189 158 124 92 90 0
197 187 195 198 185 186 115 78 81 96 1/9 1/9 1/9 0 193 193 191 186 154 122 92 89 0
194
194
190
194
198
190
193
190
187
179
177
177
88
93
86
99
94
95
69
100  j 1/9 1/9 1/9 = 0
0
194
194
192
192
189
188
184
182
149
146
121
122
91
93
90
95
0
0
201 194 191 186 186 181 74 110 82 76
196 194 195 191 183 164 77 119 84 88 1/9 1/9 1/9 0 195 193 190 183 147 128 100 106 0
0 194 192 189 181 146 125 100 105 0
192 194 199 192 191 174 89 164 103 129
i 0 0 0 0 0 0 0 0 0 0
201 190 187 189 178 168 90 82 88 84

height width
H ( x, y )    I (i, j )M ( x  i, y  j )
j 1 i 1

Write your own convolution function

myconv.m to perform a convolution.
It should accept two parameters – the
input matrix (image) and convolution
http://www.s2.chalmers.se/undergraduate/courses0203/ess060/PDFdocuments/ForScreen/Notes/Convolution.pdf
kernel, and output the filtered matrix.

15
 Convolution example in 1D
Horizontal slice from Mandrill image
Filtered Signal

1D Gaussian filter

 =

0.01 0.08 0.24 0.34 0.24 0.08 0.01

Image Processing using Matlab Sumitha Balasuriya 16

 Common convolution kernels
0.11 0.11 0.11 -0.17 -0.67 -0.17 -0.17 -0.67 -0.17 0.01 0.08 0.01
0.11 0.11 0.11 -0.67 3.33 -0.67 -0.67 4.33 -0.67 0.08 0.62 0.08
0.11 0.11 0.11 -0.17 -0.67 -0.17 -0.17 -0.67 -0.17 0.01 0.08 0.01
Arithmetic mean Laplacian (enhance edges) Sharpening filter Gaussian filter
filter (smoothing) >>fspecial('laplacian') >>fspecial('unsharp') (smoothing)
>>fspecial('average') >>fspecial('gaussian')

Investigate the listed kernels in Matlab by

1 0 -1 1 2 1 performing convolutions on the Mandrill and
2 0 -2 0 0 0 Lena images. Study the effects of different
kernel sizes (3x3, 9x9, 25x25) on the output.
1 0 -1 -1 -2 -1
Sobel operators (edge detection in x
and y directions) The median filter is used for noise reduction. It works by
>>fspecial('sobel') replacing a pixel value with the median of its neighbourhood
pixel values (vs the mean filter which uses the mean of the
>>fspecial('sobel')’
neighbourhood pixel values). Apply Matlab’s median filter
function medfilt2 on the Mandrill and Lena images. Remember
to use different filter sizes (3x3, 9x9, 16x16).
17
 Useful functions for convolution
• Generate useful filters for convolution
>>fspecial('gaussian',[kernel_height kernel_width],sigma)

• 1D convolution
>>conv(signal,filter)

• 2D convolution
>>conv2(double(I(:,:,2)),fspecial('gaussian‘,[kernel_height kernel_width] ,sigma),'valid')

image kernel Border padding options

Perform the convolution of an image using Gaussian

kernels with different sizes and standard deviations
and display the output images.
18
 Tutorial 2
1) Type the code in this handout in Matlab and investigate the results. function H=grating2d(f,phi,theta,A)
2) Do the exercises in the notes. % function to generate a 2D grating image
3) Create a grating2d.m function that generates a 2D steerable spatial frequency. % f = frequency
Compute spatial frequencies with an amplitude =1 and the following parameters % phi = phase
frequency = 1/50, 1/25, 1/10, 1/5 cycles per pixel, % theta = angle
phase= 0, pi/5, pi/4, pi/3, pi/2, pi % A = amplitude
theta = 0, pi/5, pi/4, pi/3, pi/2, pi % H=grating2d(f,phi,theta,A)
The value for pi is returned by the in-built matlab function pi. % size of grating
Display your gratings using the in-built gray colourmap. (figure 1) height=100;
3) Create a superposition of two or more gratings with different frequencies and width=100;
thetas and display the result. You can do this by simply adding the images you
generated with grating2d (figure 2) wr=2*pi*f; % angular frequency
frequency = (1/10 and 1/20), (1/20 and 1/30) wx=wr*cos(theta);
theta = (pi/2 and pi/5), (pi/10 and pi/2), (pi/2 and pi) wy=wr*sin(theta);
Make sure you examine combinations of different frequencies and theta values. for y=1:height
(figure 3). for x=1:width
Visualise the intensity surface of the outputs that you have generated. (figure 4) H(x,y)=A*cos(wx*(x)+phi+wy*(y));
4) Write a matlab function that segments a greyscale image based on a given end
threshold (i.e. display pixels values greater than the threshold value, zero end
otherwise). The function should accept two inputs, the image matrix and the
threshold value, and output the thresholded image matrix. (figure 5)

Figure 1 Figure 2 Figure 3 Figure 4 Figure 5

19