Friday, October 24, 2014

Data Compression : What Makes Data Compression Possible

Data compression is referred as the process to reduce the size of data required to represent certain amount of information. Data and information are not the similar. Data refers to the way by which the information is conveyed. Various amounts of data can symbolize the same amount of information. Every so often the given data contains some data which has no relevant information, or repeats the identified information. It means there is data redundancy in image.
 

Data redundancy is the essential concept in image compression and can be mathematically defined.
The relative data redundancy R is: R= 1-1/C
Where C= compression ratio and C= N1/N2
Where N1 is original image size and N is compressed image size.
Ø    In general, three basic redundancies exist in digital images as follows:

Psycho-visual Redundancy: It is a redundancy corresponding to different sensitivities to all image signals by human eyes. Therefore, removing few less vital information in our visual processing might be reasonable. Removing this kind of redundancy is very lossy process and the gone information can't be recovered. To remove such kind of redundancy generally used a method that is called quantization which means the mapping of a range of input values to a limited number of output values..
Inter-pixel Redundancy: This type of redundancy is associated with the inter pixel correlations within an image. Much of the visual contribution of a single pixel is redundant and can be guessed from the values of its neighbors. This type of redundancy can be removed by run-length coding.

Coding Redundancy: The uncompressed image usually is coded with each pixel by a fixed length. For example, an image with 256 gray scales is represented by an array of 8-bit integers. Using variable length code schemes such as Huffman coding and arithmetic coding may produce compression in which shortest code words assign to the most frequent grey levels and longest code words assign to the least frequent grey levels.

 
Temporal Redundancy: Temporal Redundancy is the redundancy of information which survives among set of frames. Thus it can be referred as inter frame redundancy. In video sequence, the numerical values of the luminance/brightness and chrominance/color of all the pixels in a given frame is either exactly similar or at least near similar to those values in the preceding frames. In simpler words, this redundancy is the ‘repetition of information between frames’. Motion estimation techniques are used to eliminate temporal redundancy.


                                                                         Author - Hemika Yadav
(Research Associate at Silicon Mentor)