Scribd
Upload
3 views

lec04_compression_part2

The document provides an overview of image compression techniques in digital image processing, including definitions, compression ratios, and various methods such as Huffman coding, arithmetic coding, and dictionary-based methods like LZ77 and LZW. It discusses the principles behind these techniques, including lossless and lossy compression, and the importance of data redundancy. Additionally, it covers predictive coding and run-length coding as methods for efficient data representation.

Uploaded by

Tachbir Dewan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
3 views

lec04_compression_part2

The document provides an overview of image compression techniques in digital image processing, including definitions, compression ratios, and various methods such as Huffman coding, arithmetic coding, and dictionary-based methods like LZ77 and LZW. It discusses the principles behind these techniques, including lossless and lossy compression, and the importance of data redundancy. Additionally, it covers predictive coding and run-length coding as methods for efficient data representation.

Uploaded by

Tachbir Dewan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 23

COL783: Digital Image Processing

Image Compression
Prem Kalra
[email protected]
h4p://www.cse.iitd.ac.in/~pkalra/col783

Department of Computer Science and Engineering


Indian InsEtute of Technology Delhi
Recap
Some definitions Compression techniques
• Compression ratio • Loss-less and Lossy
• Fidelity criteria • Symmetric and Asymmetric
Data Redundancy Variable length coding
• Coding • Huffman Coding
• Interpixel • Information theoretic analysis
• Psychovisual Entropy

Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783


Image Compression
Kraft’s inequality

Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783


Image Compression
Lower Bound

Shannon’s Coding Theorem

Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783


Image Compression
Lower Bound

Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783


Image Compression
Efficiency of Huffman Coding
H(z)/L(z)

Variants of Huffman Coding


• Higher order estimate of entropy
• Truncated Huffman Coding
• Dynamic or Adaptive Huffman Coding

Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783


Image Compression
Arithmetic Coding
Basic Idea:
a) Like Huffman coding requires prior knowledge of probabiliEes
b) Unlike Huffman coding, which assigns variable length codes to symbols arithmeEc
coding assigns codes to a variable group of symbols i.e. the message.
c) There is no one-to-one correspondence between the symbol and its
corresponding code word.
d) The code word itself defines a real number within the half-open interval [0,1) and
as more symbols are added, the interval is divided into smaller and smaller
subintervals, based on the probabiliEes of the added symbols.

Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783


Image Compression
Arithmetic Coding

End of message or length of message is known.


Source: Digital Image Processing, Gonzalez and Woods.
Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783
Image Compression
Arithmetic Coding

Final code 068

Source: Digital Image Processing, Gonzalez and Woods.


Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783
Image Compression
Arithmetic Decoding
Follows encoding procedure
Code 068 may be converted to the real number 0.068, which falls in the
first sub-interval [0,0.2) therefore first symbol is a1, and so on.

Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783


Image Compression
Dictionary based methods
• Compressing mulEple strings can be more efficient than
compressing single symbols only (e.g. Huffman encoding).
• Strings of symbols are added to a dicEonary. Later occurrences
are referenced.
• StaEc dicEonary: Entries are predefined and constant according
to the applicaEon of the text
• AdapEve dicEonary: Entries are taken from the text itself and
created on-the-fly

Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783


Image Compression
Dictionary based methods: LZ77
By Lempel and Ziv in 1977 about lossless compression with an adapEve
dicEonary.
• Runs through the text in a sliding window
• Two buffers are used - search (history) buffer and a look ahead buffer.
• The search buffer is used as dicEonary
• Sizes of these buffers are parameters of the design

Source: h4p://jens.jm-s.de/comp/LZ77-JensMueller.pdf
Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783
Image Compression
Dictionary based methods: LZ77
Look-ahead

Search

Source: h4p://jens.jm-s.de/comp/
LZ77-JensMueller.pdf
Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783
Image Compression
Dictionary based methods: LZ77

Decoding

Source: h4p://jens.jm-s.de/comp/
LZ77-JensMueller.pdf
Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783
Image Compression
Dictionary based methods: LZW
Extended by Welch (Lempel, Ziv and Welch)
This coding scheme has been adopted in a variety of imaging file formats, such as
the graphic interchange format (GIF), tagged image file format (TIFF) and the
portable document format (PDF).

Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783


Image Compression
Dictionary based methods: LZW
Extended by Welch (Lempel, Ziv and Welch)
• Unlike Huffman coding and arithmeEc coding, this coding scheme does
not require a priori knowledge of the probabiliEes of the source
symbols.
• The coding is based on a “dicEonary” or “codebook” containing the
source symbols to be encoded. The coding starts with an iniEal
dicEonary, which is enlarged with the arrival of new symbol sequences.
• There is no need to transmit the dicEonary from the encoder to the
decoder. The decoder builds an idenEcal dicEonary during the decoding
process
Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783
Image Compression
Dictionary based methods: LZW
Extended by Welch (Lempel, Ziv and Welch)
Example: 32 32 34 32 34 32 32 33 32 32 32 34

Consider a dicEonary of size 256 locaEons (numbered 0 to 255) that contains


entries corresponding to each pixel intensity value in the range 0-255.

Source: h4ps://nptel.ac.in/courses/117/105/117105083/#
Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783
Image Compression
Dictionary based methods: LZW
Extended by Welch (Lempel, Ziv and Welch)

Source: h4ps://
nptel.ac.in/courses/
117/105/117105083/#
Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783
Image Compression
Run Length Coding
Run: a string of the same symbol

Example
input: AAABBCCCCCCCCCAA
output: A3B2C9A2

compression ratio = 16/8 = 2


Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783
Image Compression
Predictive Coding
Basic premise: Current pixel is similar to the previous pixel
(coherence)
Differential Coding
d(x,y) = I(x,y) – I(x-1,y)
d(x,y) prediction error which is to be encoded.

Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783


Image Compression
Predictive Coding
Compression

Source: Digital Image Processing, Gonzalez and Woods.


Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783
Image Compression
Predictive Coding
Decompression

Source: Digital Image Processing, Gonzalez and Woods.


Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783
Image Compression
Predictive Coding

Digital Image Processing http://www.cse.iitd.ac.in/~pkalra/col783

You might also like