Jun 21 • Notes • 2201 Views • 2 Comments on IMAGE COMPRESSION


When the large amount of data is stored in a space is known as image compression.

It is of two types

  1. Lossless compression
  2. Lossy compression


  • It is a reversible process
  • No information is lost
  • It is useful for text data which is easily understandable by human.
  • This compression is independent of the psychovisual system
  • It is useful in the domain where reliability is very crucial.

There are some lossless algorithm and these are

  1. Run-length coding
  2. Huffman coding
  3. Dictionary based or LZW coding
  4. Golomb coding
  5. Symbol based coding


With the help of Run-length coding length of pixel value is identified and the images is encoded in the form of RUN. Each row of image is written in the form of sequence. Run of black or white pixel represents the length of row. This is known as run-length coding.

Standard that is used to encode the binary and grey-level image is CCITT(consultative committee of the international telegraph and telephone) and run-length coding is CCITT. By this technique images are scan row by row and identifies the run. Pixel value and length of the run is specified by output run-length vector.

Hrun-length = ( H0 + H1)/ (L0 +L1)

H0 = entropy of the black run

H1 = entropy of white run

L0 = average value of black run

L1 = average value of white run


It is a variable length coding type. By selecting the better way of assigning the codes coding redundancy is eliminated in Huffman coding.

Algorithm for Huffman coding

  1. List all the symbols and arrange them
  2. Select two symbols having least probability
  3. Create a node and add the probability of selected symbols and make anew node.
  4. Repeat the process till only one node remains
  5. Then assign  the code 0 for left and 1 for another
  6. Code from the root to leaf is traced.


LZW means Lempel-Ziv-Welch. And this coding is use to store the string pattern in dictionary that have already been encountered repeated patterns are encoded by the indices. And the encoders read the string and then recurrent word is identified and gives the indices of word from the dictionary. And if new word is encountered then it is entered into the dictionary as a new entry. It is also known as dictionary based coding.

Advantage of LZW coding:

  • This coding is faster
  • LZW coding is not based on statistics thus they are independent of quality of the model on data distribution.
  • And these are adaptive in the nature.


This coding is invented by Solomon W. Golomb in 1960s. input value is divided into two part for this golomb coding uses uses tunable parameter M. q is the quotient and r is remainder and then quotient is send in unary coding and the remainder in truncated binary coding. Golomb coding is equivalent to unary coding when M = 1.

When occurrence of small values in the input stream is significantly more likely than large values in this situation golomb coding is highly suitable.


It is also known as the token-based coding. Symbol is a collection of frequently occurring sub images. In a symbol dictionary each symbol is stored and image is coded as asset of triplet {(x1, y1, t1), (x2, y2, t2)……….} where location of symbol is specifies by the (xi, yi) and ti is token is a address of the symbol in dictionary an instance of dictionary symbol in image is represented by each triplet.


The term digital watermarking is first given by Andrew Tirkel and Charles Osborne.

Ownership of the copyright of noise-tolerant signal is identified by the use of digital watermarking. Digital information in a carrier signal is hide by the process of watermarking. This process is also used to verify the integrity or authenticity of carrier signal. Mainly it is used for the tracing copyright infringements and for banknote authentication. only under certain condition digital watermarking  is predictable.

Digital watermarking life cycle phase:




  1. What is image compression???


When large amount of data is stored in a space is called as image compression.

2. What are the type of image compression??


There are two type of image compression

a.Lossless compression

b. Lossy compression

3.Explain lossy compression???


Lossy compression

  • It is a non reversible process
  • In this compression information is lost
  • Compression ratio is very high
  • It is useful for the type of data which is not easily understandable or integrated by the human.
  • It is requiredin the domain where loss of data is acceptable.                                                            4. .What are the lossless compression algorithm????


There are five types of coding

  1. Run-length coding
  2. Huffman coding
  3. LZW coding
  4. Golomb coding
  5. Symbol based codiing                                                                                                                                              5. What is use of digital watermarking??


Use of digital watermarking

  • Verify the integrity and authenticity of carrier signal
  • For tracing copyright infringements
  • For authentication of banknote

Tell us Your Queries, Suggestions and Feedback

Your email address will not be published.


  1. priyanka nayak says:

    This article clearly describes about image processing which comes under the signal processing unit. The concept described here is really helpful

  2. Surabhi Singh says:

    This article carries all the informations related with Image compression. a very useful article to check out to enhance knowledge.

« »