Color Image Processing Image Compression

Topics
Color Image Processing Color Models RGB YIQ HSI Image Compression Lossy and lossless methods Run length coding Entropy coding Arithmetic coding Jpeg coding

Color Image Processing: Motivation

Color is an important feature for detecting and

identifying objects
Human face detection License plate detection Content based image retrieval systems

Image segmentation can be performed base on color

values Tracking moving objects in video (surveillance ) Etc.

Color Spectrum

Visible Wavelength

Absorption of light by human eye

Red, Green, Blue Color Cube

RGB Color Cube

YIQ Color Model

YIQ is the color space used by the NTSC color TV

system The Y component represents the luma information, and is the only component used by black-and-white television receivers I and Q represent the chrominance information For example, applying a histogram equalization to a color image is done by Y component only

YIQ Color Model Conversion

HSI (or HSV) Color Model

Conversion from RGB to HSI

Enhancement using HSI Model

Color Space Conversions in MATLAB (1)

To convert from RGB to HSV use rgb2hsv
For HSV to RGB conversion use hsv2rgb E.g. I = imread(test.bmp); H = rgb2hsv(I);

Color Space Conversions in MATLAB (2)

To convert from RGB color space into YIQ color space

use rgb2ntsc
RGB = imread(sample.png'); YIQ = rgb2ntsc(RGB);

ntsc2rgb convert from YIQ to RGB Images can be displayed in RGB space only

Color Space Conversions in MATLAB (3)

I component in YIQ (the first component) is

equivalent to a gray scale conversion. However, you may also use rgb2gray either.
I=imread(test.bmp); YIQ = rgb2ntsc(I); imshow( YIQ(:,:,1))

Image Compression

Image Compression
Importance of Image Compression Typical image resolution/depth

1024 x 1024 x 24 Typical image size in bytes: 3 MB 30 fps, 720576 , 24 bits Typical video size for 1 minute: More than 2 GB

Typical video rate/resolution/depth

Data Redundancy
Redundancy in information theory is the number of

bits used to transmit a message minus the number of bits of actual information in the message.
Informally, it is the amount of wasted "space" used to

transmit certain data

Data Compression
Data compression or source coding is the process of

encoding information using fewer bits (or other information-bearing units) than an un-encoded representation would use through use of specific encoding schemes.

Compression Methods
Lossless Methods Lossy Methods

Run Length Coding

In this method runs of data (that is, sequences in which the

same data value occurs in many consecutive data elements) are stored as a single data value and count. Useful in relatively simple graphic images such as icons, line drawings, and animations. Example, WWWWWWWWWWWWBWWWWWWWWWW WWBBBWWWWWWWWWWWWWWWWWW WWWWWWBWWWWWWWWWWWWWW is stored as : 12W1B12W3B24W1B14W

Entropy Coding
Entropy coding is a lossless coding method. Entropy coding creates variable length codeword. The

length of each codeword is approximately proportional to the negative logarithm of its probability. Huffman Coding Huffman coding is an entropy encoding algorithm used for lossless data compression. The term Entropy Coding refers to the use of a variable-length code table for encoding a source symbol

Example of Variable Length Coding

Huffman Coding Example

Huffman Coding Example

Arithmetic Coding
The output from an arithmetic coding process is a single

number less than 1 and greater than or equal to 0

This single number can be uniquely decoded to create

the exact stream of symbols that went into its construction.

Example: a1a2a3a3a4

Example

Lossy Compressions
A lossy compression is a data compression method

which discards (loses) some of the data, in order to achieve a high compression rate, with the result that decompressing the data yields content that is different from the original one.

JPEG Standard
Steps followed in JPEG Color space transformation Down-sampling Block splitting Discrete cosine transform Quantization Run length and Entropy coding

Color space transformation

First, the image should be converted from RGB into

YCbCr color space The compression is more efficient as the brightness information, which is more important to the eventual perceptual quality of the image, is confined to a single channel, more closely representing the human visual system.

Down Sampling
Due to the densities of color- and brightness sensitive

receptors in the human eye, humans can see considerably more fine detail in the brightness of an image (the Y component) than in the color of an image (the Cb and Cr components). Down-sample Cb and Cr components only 4:2:0

Splitting the Frame into Blocks

After sub-sampling, each channel must be split into

88 blocks of pixels.

Discrete cosine transform

Discrete Cosine Transform

Result of DCT

Quantization
Quantization is done by simply dividing each

component in the frequency domain by a constant for that component, and then rounding to the nearest integer

Sample Quantization Table

Quantization Results

Zig-Zag Scanning and Run Length Coding

Zig-Zag Scanning

Variable length codes

Variable length codes are generated for the code words

of the zig-zag scan result. Trailing zeros are truncated by inserting an EOB symbol after the last non-zero value.

Typical compression ratios

The compression ration depends on the quantization

factors used. Typical ratio is 4:1

Questions?