Original tutorial written by Keith Yerex

Getting Started

Some useful commands to get aquainted with:

help (or helpwin)	The help command is like the man page in unix. Executing help will give you all the possible primary topics. You can use help topic(ie. help general) to get a listing of the help for the topic. If you know the name of a function that you wish to get help from, use help functioname. Try "help images" to obtain a list of functions in the image processing toolbox.
lookfor	Looks for the provided string in the first like of the help files of every fill in the MATLABPATH. Useful if you know a partial name of a function.

Try entering 3+7 at the command line. You will notice that it displays the result. If you place a semicolon at the end of the line, then the result is not displayed 3+7;. You may want to assign a variable, so x=3+7; and not want to see the result immediately. Now to display the contents of x, you can either type x, or disp(x). For the above example when you typed 3+7; you did not get a result displayed, but matlab keeps this answer in a variable ans, which you can use in the next expression, etc. Keep in mind that ans gets reassigned every time you evaluate an expression without assigning it to a variable.

Matlab has a current workspace, which keeps track of all the variables that have been assigned. You can see what variables are currently assigned by issueing a whos command. This displays the type,name and size of all the current variables. You can issue a whos x to get information about x only. You can also clear the entire workspace, or delete a variable by calling clear or clear variablename.

You can also preserve the current workspace for later use, by using the save command. The save command will write a file called matlab.mat, while save myfile saves to myfile.mat. The previous calls save the entire workspace, to save the state of a variable(s) just call save xdata x(additional variables can be saved by placing leaving replacing x in the previous command to x1 x2 etc.) to save the variable x to the file xdata(.mat files can get very large when you are dealing with images). You can load these .mat files at a later time by using the load command, which has similar syntax to the save command.

clear;
whos;
a=1;
b=a*a+2;
s='string';
whos;
save removethisfile a b s
clear;
whos;
load removethisfile;
whos;

v=[1,2,3];
w=[3,2,1];
x=v+w;

Another way to initialize vectors is to use the colon notation. The colon notation is first:step:last,where x=2:1:5 implies that x=[2,3,4,5]. The step is optional, if it is not there the stepping is done in increments of 1. You can also use the colon notation to access some elements of an array, or all of them. v(2:3) is the list [2,3] or v(1:3) returns [1,2,3]. Note that although v and w are row vectors v(1:3) returns a row vector, but v(:) returns a column vector.

Similar to the vectors matrices are indexed starting from 1. So to access element at the I-th row and J-th column from matrix A, you have A(i,j).

Matrix initialization is done by specifying each row separated by semicolons or spaces. So :

A=[1,2,3;4,5,6;7,8,9];
gives us the matrix:
1 2 3
4 5 6
7 8 9

You can assign an entire row using:

A(1,:)=[1,2,3];

A(:,1)=[1,4,8];

To determine the size of an vector or matrix you can use size(A), which will return a row vector containing the dimensions of A. To determine the size of the nth dimension you can use size(A,n).

Multiplication         *
Tranpose               '
pointwise mult        .*
pointwise div         ./
inverse               inv
determinant           det
cross            cross product of 2 vectors
Linear Least Squares x=A\b (for the system Ax=b)
...

A = [1,2,3;4,5,6];
B = A+1;
C = B.*A; % point wise multiplication
A = rand(100,3);
b = [1:100]';
x = A\b; % linear least squares solution to Ax=b

Matlab uses figures to display graphical information. A figure is basically a window, and if you would like to pop up a window, use figure. You can use figure to pop up several windows, if you give it a handle as an argument(ie, figure(1)). clf clears the current figure, and gcf gets the current figure.
The plot function can be used to plot data in a 2d format. The hold on/off can be used to plot different data to the same axis.

figure(1);
x=[1:10]
plot(x,x.*x,'r');
xlabel('x-label');
ylabel('y-label');
hold on
plot(x,x.*x.*x,'b');
title('x^2 in red, x^3 in blue');
hold off;

Plotting some surfaces

[x,y]=meshgrid(-1:0.2:1,-1:0.2:1);
z = power(x+y,2);
subplot(1,3,1);
mesh(x,y,z);

[x,y]=meshgrid(-1:0.2:1,-1:0.2:1);
z2= x+y;
subplot(1,3,2);
mesh(x,y,z2);

subplot(1,3,3);
mesh(x,y,sin(4*x)+cos(4*y))

Many control flow expressions the same as c/c++/java

You can use these operators on entire matrices or vectors, and there will be ones and zeros depending on the output of the operator.

A = [1,2,3];
B = [3,2,1];
A==B
B~=A

if (condition)
statments...
elseif (condition2)
statments..
...
end

for index=first:step:last
   statments
end

while condition
   statments
end

Commands in a script are run as if they are typed at the keyboard, so any variables you have currently been using in your workspace can be used in the script. Also, any variables created in the script are available after the script has executed.

nargin is the variable that contains how many variables were sent to your function. nargout is variable that contains how many variables are awaiting to be bound from your function. Sample function

function [outputargs] = fname(inputargs)
% Comments that appear in help file when help fname is called
%
%
statements in the body...

Write a simple function computing the factorial of a given input.

imread is used to load an image from a file. The image that you read in can be displayed using both imshow or image. When an rgb image is read in, the third dimension is the color data, so im(:,:,1) is the red channel, im(:,:,2) is green, and im(:,:,3) is blue.
Note: images are usually stored in unsigned chars(bytes) which is a uint8 type in matlab. You can not perform many operations on these data types so you must convert (like a cast) into some other appropriate type(double, uint8, uint16, uint32, int16, int32)

im = imread('image.jpg');
imshow(im);       % or image(im);
im = im-im/2;     % should be an error, since image is probably uint8
im = double(im);  % convert to double
im = im+10;       % this is now ok
imshow(uint8(im));

im=imread('image.jpg'); % show default image of a child
P=ginput(2);% waits for two button presses
hist(im(:),0:255) %note this change

Using complex Numbers for 2d points. Assuming a point in 2d is represented as p=[x,y], we can use complex numbers to store and operate on this number.

  c = x+yi, where i=sqrt(-1);
 

   r = cos(theta) + sin(theta)*i;
 

• make sure you don't overwrite i and j as variables(possibly in a loop) and then try using i or j as sqrt(-1). I do not mean "never" use i and j as variable names, I am just saying be aware that if you do overwrite them and attempt to use them later as if they were sqrt(-1)

Additional information on text and file input/output, structures, cell arrays, improving matlab code, see the tutorial or the help files.