# Basic manipulation of images in MATLAB

The objective of this part is to be familiarized with basic read/write operations on images in MATLAB, and learn the type of images supported as well as some basic operations over them.

## Basics

Images have a size in pixels, that we usually express as NxM. In MATLAB, every image is nothing but a matrix. Each element of the matrix is the gray tone for images in B/W, or, if it is a color image, it can be represented by three matrixes, where each matrix represents one of the three primary colors (red, green, blue or RGB).

To read an image from disk, we can use imread:

>> f = imread('chestxray.jpg')
>> f = imread('/home/user/chestxray.jpg')

If the image is B/W, MATLAB will create a matrix f, where each element will have a value of gray:

$f=\left[\begin{array}{cccc}f\left(1,1\right)& f\left(1,2\right)& \cdots & f\left(1,N\right)\\ f\left(2,1\right)& f\left(2,2\right)& \cdots & f\left(2,N\right)\\ ⋮& ⋮& & ⋮\\ f\left(M,1\right)& f\left(M,2\right)& \cdots & f\left(M,N\right)\end{array}\right]$

Now we can check the size of the image:

>>size(f)
ans =
1024 1024

We can save the size in a variable:

>>[M, N] = size(f)

We could also save the number of colors of the image with the same command:

>>[M, N, c] = size(f)

And if we want detailed information about the image:

>>whos f
Name Size       Bytes    Class
f    249x500    373500   uint8 array
Grand total is 373500 elements using 373500 bytes

## Displaying images

We may want to display our image or apply effects to it. We can do that with imshow():

imshow(f,G)

f is the image and G is the number of intensity levels to show. If G is not specified, 256 is used. The syntax is:

imshow(f,[low high])

For example:

imshow(f) imshow(f,[0,50])

If we want to expand the dynamic range, we can leave the brackets empty. That will use the minimum intensity value of the image as the low limit, and the maximum value as the high limit:

imshow(f) imshow(f,[])

Another interesting function is pixval, which affects the last loaded image. When you run it in the console, you can see the value of the intensity when the mouse hovers a particular pixel. It can also measure the euclidean distance between two points.

Finally, when a new image is loaded, MATLAB overwrites the window of the previously loaded image. To avoid this, and open a new image in an independent window, do:

>> figure, imshow(f)

## Saving images

To save an image:

>> imwrite(f, 'filename')

If 'filename' doesn't have an extension that MATLAB recognizes, use:

>> imwrite(f, 'filename', 'tif')

We'll be working mainly with JPEG files, in which case you can also specify the quality:

>> imwrite(f, 'filename.jpg', 'quality', q)

Where q is a number between 0 and 100 (0 = biggest compression, worse quality, 100 = smallest compression, best quality).

To get information about an image:

>> imfinfo filename

would show:

>> imfinfo test.jpg
ans =
Filename: 'test.jpg'
FileModDate: '08-feb-2005 17:18:13'
FileSize: 6125
Format: 'jpg'
FormatVersion: ''
Width: 600
Height: 494
BitDepth: 8
ColorType: 'grayscale'
FormatSignature: ''
NumberOfSamples: 1
CodingMethod: 'Huffman'
CodingProcess: 'Sequential'
Comment: {}

## Image types

double
Double precision, numbers in floating point that vary in a range of -10308 to 10308 (8 bytes per element)
uint8
8 bit integers in the range of [0,255] (1 byte per element)
uint16
Integers of 16 bits in the range of [0, 65535] (2 bytes per element)
uint32
Integers of 32 bits in the range of [0, 4294967295] (4 bytes per element)
int8
Integers of 8 bits in the range of [-128, 127] (1 byte per element)
int16
Integers of 16 bits in the range of [-32768, 32767] (2 bytes per element)
int32
Integers of 32 bits in the range of [-2147483648,2147483647] (4 bytes per element)
single
Floating point number of simple precision, with values in the range of -1038 to 1038 (4 bytes per element)
char
Characters (2 bytes per element)
logical
The values are 0 or 1 (1 byte per element)
Intensity images
A matrix whose values have been scaled to represent intensity. They could be uint8 or uint16. If they are double, the values are scaled in the range [0,1].
Binary images
Images whose values are only 0 or 1. They are represented in MATLAB using logical values. To convert an array of zeros and ones into a logical array in MATLAB, we type:
>> B = logical(A)
To check if an array is logical:
>> isLogical(A)

## Converting images

Image type conversion table
Command Converts to Valid entry type
im2uint8 uint8 logical, uint8, uint16 and double
im2uint16 uint16 logical, uint8, uint16 and double
mat2gray double ([0,1]) double
im2double double logical, uint8, uint16 and double
im2double logical uint8, uint16 and double

For example:

>> f = [0 0.5; 0.75 1.5]
f =
0		0.5000
0.7500	1.5000

>> g = im2uint8(f)
g =
0	128
191	255
>> g = [0 0.3; 0.7 0.9]
g =
0		0.3000
0.7000	0.9000

>> gb = im2bw(g, 0.6)
gb =
0	0
1	1

You can practice these concepts using snippet01.