Image Representation
LECTURE-2
Digital Image
Virtually all computing devices in use today are digital.
Data is representation in a discrete form using patterns of binary
digits (bits) that can encode numbers within finite ranges and with
limited precision.
The images we perceive in our environment are analogue and
formed by the complex interaction between light and physical
objects.
Suppose we wish to capture an image and represent it in a computer.
Since we do not have infinite storage (bits), it follows that we must
convert that analogue signal into more limited digital form. This
conversion process is called sampling.
That means the process of getting finite number of values from image (cont. func.)
is called sampling.
Raster Image Representation
The problem of digital image representation is that it can
not encode fine levels of details, nor dynamic color
ranges.
The solution of this problem is to break the image up into
a regular grid that we call raster.
A raster image, stored in the form of colored pixels on
a fixed and finite grid.
Each grid cell is “Picture Cell” often known as pixel.
Smallest element of the screen.
Atomic unit of the image.
Pixels are arranged in a row column fashion to form a
rectangular picture area (raster).
Height (H)
Raster Image Representation
Total number of pixels in an image is a function of size of
the image and resolution.
3x2 inch image at a resolution of 300 pixels per inch would have total
pixel= (3*300)* (2*300)=540000 pixel.
Number of pixels per unit length (e.g. inch) in the
horizontal as well as vertical direction is known as
resolution.
Image size is given by-
Total number of pixels in the horizontal direction x Total
number of pixels in the vertical direction
Example: 512 x 512 or 640 x 480
Aspect Ration= Width/Height
[measured in unit length or number of pixels]
2 x2 inch image and 512 x 512 image have aspect ratio of 1/1
Raster Image Representation
Individual pixels in an image can be referenced by their
coordinates.
Left Handed System:
Origin->top left corner
(W-1, H-1) pixel->bottom right corner.
Right Handed System:
Origin->lower left corner
(W-1, H-1) pixel->upper right corner
Raster images are pixel based and format is resolution
specific (Better resolution->better quality).
Common raster format: TIFF, JPEG, GIF, PNG and BMP
Examples
Compute the size of a 640 x 480 image at 240 pixels
per inch.
(640/240) x (480/240)= 2 2/3 x 2 inch
Compute the resolution of a 2 x 2 inch image that has
512 x 512 pixels.
512/2 = 256 pixels per inch
If an image has a height of 2 inches and aspect ratio
of 1.5, what is its width?
Aspect Ratio= Width/Height
Width=Aspect Ratio x Height= 1.5 x 2= 3 inches
Problems
If we want to resize a 1024 x 768 image to one that is 640
pixels wide with the same aspect ratio, what would be the
height of the resized image?
If we want to cut a 512 x 512 sub image out from the
center of an 800 x 600 image, what are the coordinates
of the pixel in the large image that is at the lower left
corner of the small image (Right Handed System)?
How to convert the coordinate of a pixel at (x, y) in the
left handed coordinate system into its coordinate (x´, y´)
in the lower left corner as origin coordinate system (right
handed system)?
Vector Graphics Representation
Based on the mathematical formulas that define
geometric primitives such as polygons, lines, curves,
circles & rectangles.
Stored in the form of mathematical representation of
the image element.
Points, lines and curves may be scaled up or down to
any resolution with no aliasing.
The points determine the direction of vector path.
Each path may have properties including values for stroke
color, shape, curve, thickness and fill.
Rasterization & Vectorization
Rasterization:
Process of converting vector graphics format to raster
graphics.
Vectorization:
Process of converting raster graphics to vector graphics format
Hardware Frame Buffer
A large piece of memory reserved for holding the bit
mapped image (color of each pixel) that is
continuously being sent to the display device is called
frame buffer.
Frame buffer is stored in main memory.
Grayscale Frame Buffer
Simplest form of frame buffer.
Encodes pixels using various shades of gray.
Pixels are encoded as an unsigned integer using 8 bit
(1 byte) and can present 256 (28) different shades.
0 means black, 255 means white and mid intensity
grey is 128.
Image of width W pixels & height H pixels requires
W x H bytes of memory for its frame buffer.
x
➢Frame buffer is arranged so that the first byte of memory corresponds to the
pixel at coordinates (0,0) [left handed system].
➢Addressing then proceed in a left to right and then top to down manner.
➢Pixel (x, y) would be stored at buffer offset A where
A= x + W y
A bytes from the start of the frame buffer.
➢Sometimes, we use the term scanline to refer to a full row of pixels. A scanline is
therefore W pixels wide.
True Color Frame Buffer
Represents color images.
RGB color value for each pixel is stored directly within
the frame buffer.
Each pixel requires at least 3 bytes. One byte for each
primary color.
Possible variation of color 224 (3 byte=24 bit)
Require 24 bit of storage per pixel.
Pixel (0,0)->Buffer location 0,1,2
Pixel (1,0)-> Buffer location 3,4,5
So, Pixel (x,y)-> S=3W
A= 3x+Sy
Where S is the stride of the display.
Representation of Color
Red, Green and Blue are called “additive primaries”
because we obtain other, non primary color by
blending (adding) different quantities of Red, Green
and Blue light.
We can make the additive secondary colors by
making pairs of primaries: Red, Green and Blue.
Red + Green=Yellow
Green + Blue= Cyan
Red + Green +Blue = White
No mix of RGB= Black
Additive Vs Subtractive Primaries
Additive color model are used in light.
Start with black and add primary component to make desired
shade.
Subtractive color models are used with paint.
Start with white and take away appropriate primary
components to yield a desired color
Example: Subtract Red from White, remains Green + Blue
which is Cyan.
RGB Color Space
Each primary color can take intensity value ranging
from 0 (lowest) to 1 (highest). [R,G,B values divided
by 255 to change to o..1]
Additive color model
R + G +B->white
(0,0,0)->Black , (1,1,1)->white
(1,1,0)->Yellow, (0.7,0.7,0.7)-> Gray
Used in display monitor.
Color Cube
(1,1,1) (1,1,1)
(0,0,0) (0,0,0)
CMYK Color Space
Complementary color model of RGB.
Subtractive color model.
C=W-R, M=W-G, Y=W-B
Ink reduces the light that would otherwise be reflected. Ink
subtracts brightness from white.
K= min (c, y, m) , K for key which is a black color for printing
purpose.
Used in printer.
Formula for Conversion
Suppose, RGB value of violet is
(238,130,238)or
(0.933,0.509,0.933).
Now we want to convert this
value to CMY.
C=1-R=1-0.933=0.067
M=1-G=1-0.509=0.491
Y=1-B=1-0.933=0.067
Self Study
Grayscale Conversion
