AM 37 Introductory Image Processing

1

• AM 37 Introductory Image Processing
• Lectures 1-2
• Contents
• Image representation
• Pixels
• Binary, grey tone, and colour images
• Brightness histograms
• Reduction of colour depth
• Image compression methods
• Loss free methods
• Lossy methods

2

• A pixel The atom of a digitised image
• Usual shape Quadratic or almost quadratic
• Properties (other than position) a colour or a
  grey tone
• 1 bit per pixel usually black (0), white (1)
• 8 bits per pixel grey tone image usually black
  (0),..white (255)
• 8 bits per pixel colour image a palette
  specifies the red, green, blue intensities of the
  256 colour indices
• 24 bits per pixel First byte, second byte,
  third byte specify the red, green and blue
  component, respectively. Number of colours over
  16 million

3
Colour components
256 x 256 x 256 ? 17 mill. colors
6 x 6 x 6 216 colors RGB cubes
RGB triangle

4
Components of coloured pixels Red r Hue
h Green g Saturation s Blue b Intensity i

5


6

• A digitized image
• A two dimensional matrix of pixels
• Today
• High resolution image 2000 x 2000 pixels
• Low resolution image 512 x 512 pixels
• Representations
• On a visual medium (a PC-screen, a projector
  board, printed paper)
• On a file (disk, CD-ROM, flash memory etc.)
• uncompressed (bmp-format)
• compressed (jpg-format)
• In the RAM of a computer

7
Grey tone landscape The grey tone is the third
dimension Below the grey tone landscape is shown
at right for the upper right rectangle of the
grey tone image.
8
An image in the RAM 2D matrix of pixels In low
level language (such as C) Data structure for
example unsigned long Pix In OO-languages
(C, Java) Out The function Img.GetPixel(x,y)
returns the colour index at (x,y) of the image
Img. In The procedure Img.SetPixel(x,y,c) sets
the colour index c to the pixel at (x,y) in image
Img.
9
Brightness histograms Statistics of grey tones or
colours

• In case of colour images
• Histograms of components such as
• Red, green, blue
• Hue, saturation, intensity

10

• Reduction of colour depth
• 1) Transformation to a binary image
• a) For segmentation foreground pixels one
  colour, background pixels the other colour.
  Definitions of foreground
• Above a threshold in grey tone images
• Box in rgb-space, hue-window etc.
• b) For printing with a printer with fewer colours
  than present in the image.
• 2)Transformation from colour images to grey tone
  images
• a) Split to red, green, an blue component
• b) Split to hue, saturation, intensity
• __________________________________________________
  __________________________________________________
  _______________
• 1a) will be discussed in Lecture 3-4
• 2a) is trivial, 2b) will not be discussed further

11
1b) Reduction of colour depth for
printing Example A grey tone image with
brightness levels 0..255 is to be printed on a
printer with only two grey tones Black and
white. Method Average density over an area of
black dots is proportional to the average of
255 - the brightness value in old image Often the
image is enlarged so that a pixel in the grey
tone image contains many binary pixels
12
Best result is obtained by a method called error
diffusion (Floyd-Steinberg-method)
Special algorithms should be applied to the edges
13
Transformation using error diffusion
14

• Image compression
• Loss-free compression
• Two so called entropy coding methods
• Hufman coding
• Arithmetic coding
• Runlength coding, predictor base coding, etc.
• Example Give each word in a language a numerical
  
• code, the most frequently used words short codes.
  Then
• a text can be compressed to about half the size
  compared
• to the representation with 7-8 bits per
  character.
• Lossy (cheat-the-eye) compression
• 8x8 discrete cosine transform (old JPEG )
• Wavelet transform (new JPEG)
• Extension to motion pictures MPEG
• ) Joint Photographic Expert Group

15
The forward discrete cosine transform
The inverse discrete cosine transform
In JPEG N8
16

17

18