SAMRAT ASHOK TECHNOLOGICAL INSTITUTE,

VIDISHA

A
Synopsis
on

A novel algorithm for 2D FFT and its

inverse for Image compression

Guided By:

Submitted By:

Prof. K.G. Kirar

EI Dept. SATI

Gaurav Sharma
MTech EC IIIsem

Contents
1.
2.
3.
4.
5.

Introduction
Proposed work
Literature review
Tools required
References

Introduction
IMAGE COMPRESSION
In current era the use of digital images in various applications has
increased very rapidly. The reason behind the increased use of digital
images is the storage, transmission, modification of these images. The
video and television transmission has also become digital and this arises
the use of more and more digital images in multimedia applications.
Digital image may be seen as the rectangular array of pixels called Bitmap.
Each pixel is a small dot which represents the color (or the gray level for
grayscale images) at the particular point on the image. For the
transmission of any image the value of color at each pixel is measured and
digitally approximated is done. From this approximation, a copy of original
image is reconstructed at the receiver side. The digital image so obtained
after the approximation of the color value of each pixel in the image
requires a large space for storage purpose. For example, a moderate
image size of 512512 pixels requires 0.75 MB of disk space. One second
of digital PAL signal requires 27 MB of disk space. To store and transmit
these images compression techniques are needed. The basic aim of image
compression is to achieve low bit rate representation, requiring less space
and achieving high visual quality of decompressed image.

TYPES OF IMAGE COMPRESSION

There are two types of image compression techniques, namely, Lossless
Image Compression and Lossy Image compression. Lossless image
compression techniques are used where the reconstructed image must be
the exactly same as the original image. These kinds of techniques encode
all the information available with the original image, so that at the time of
reconstruction of image the exact replica of original image may be
obtained. Examples of lossless image compression techniques are Run
length encoding, Huffman encoding, Arithmetic encoding, Entropy coding
and Area coding etc. In case of lossy compression techniques the
reconstructed image so obtained may differ in quality compared to the
original image. Lossy compression gives higher compression ratio
compared with the lossless compression techniques. Lossy compression
techniques encode only that information which is required for the
reconstruction of the image and omits all other unnecessary information.
Though lossy compression provides higher compression ration but the
reconstructed image somewhere differ from the original image. Lossy
compression techniques are more suitable for still images. Examples of
lossy compression techniques are Predictive coding, Transform coding
(FT/DCT/Wavelets), Chroma sub sampling, Fractal compression and vector
Quantization etc.

PERFORMANCE INDICATORS:
Lossy compression techniques leave the reconstructed image with some
distortions, as the reconstructed image is only the approximation to the
original image. In order to measure and quantify the performance of
compression technique, some performance indicators are used as follows:

Compression Ratio: Compression ratio defines the ratio of number

of bits required to represent original image to the number of bits
required to represent compressed image. As the compression ratio
increases, the quality is compromised. Lossy compression techniques
have higher compression ratio than lossless compression techniques.
Mean Square Error (MSE): MSE is the measure of error between the
original image and the compressed image. Mean Square Error is the
cumulative squared error between the compressed image and the
original image. For the lesser distortion and high output quality, the
MSE must be as low as possible. Mean Square Error may be
calculated using following expression:

 Peak Signal to Noise Ratio (PSNR): PSNR is the ratio of maximum

power of the signal and the power of unnecessary distorting noise.
Here the signal is the original image and the noise is the error in
reconstruction. For a better compression the PSNR must be high. The
increase in the peak signal to noise ratio results in the decrease in
compression ratio. Therefore a balance must be obtained between
the compression ratio and peak signal to noise ratio for the effective
compression. The peak signal to noise ratio may be calculated as:

Proposed Work
Proposed Regular, FFT Algorithm for Image Processing
Twiddle Factor is a key component in FFT/IFFT computation. It is any of the
trigonometric constant coefficients that are multiplied by the data in the
course of the FFT algorithm. They are the coefficients used to combine
results from a previous stage to form inputs to the next stage.
Twiddle factor term was apparently coined by Gentleman and Sande in
1963. It is the root of unity complex multiplications constants in the
butterfly operations and also it describes the rotating vector which rotates
in increments according to the number of samples.
The twiddle factor is expressed as

By Euler's formula, the exponential term can also be written as

The irregularities of the twiddle factors in FFT can also be overcome by the
Cosine and Sine transforms of the signal. The FFT for 2D signal is obtained
by adding the Sine and Cosine transforms of the input signal. The
transforms in this algorithm is obtained in matrix form by varying the

values of u and x from 0 to 7. The transposed Sine and Cosine matrices are
obtained by varying the values of v and y from 0 to 7.
Proposed Regular FFT/IFFT Algorithm
The proposed FFT/IFFT algorithm is also folIows:
1. Read the input file from the host system.
2. Evaluate the (8x8) Cosine and Sine matrix using Equations
(

(( )

))

S(

(( )

))

3. The image is accessed as (8x8) block successively, and the Cosine and
Sine transforms are obtained.
4. Compute the FFT of the image using eqn
(

{ (

5. Verify the obtained FFT values with buiIt in FFT functional values of
MATLAB.
6. Compute the IFFT of the processed image in step 4 using the Eq.

{ (

7. Verify visually the reconstructed image with the original image.

8. Calculate the Power Signal to Noise Ratio (PSNR) of the processed
signals for validation. Note: A PSNR value of 35 dB and above implies
the reconstructed image is indistinguishable from the original image.
8

Literature review
Anitha and Ramachandran, proposed a new algorithm for image
compression using 2D FFT and its inverse and attains 35 dB of PSNR,
Their research work was published in IEEE 2013.
Sonal and Dinesh Kumar published a survey report on A Study of
various Image Compression

which explains about various

compression techniques for various application IEEE pp 63-69, 2009

Tools required
Image Processing Toolbox in MATLAB
Digital Signal Processing in MATLAB
FFT and IFFT in MATLAB

10

References

[1] Alan V. Oppenheim, Ronald W. Schafer, John R. Buck, Discrete -Time

Signal Processing, Prentice Hall, Second Edition, pp. 646-652, 1999.
[2] L. Rabiner B. Gold, Theory and Application of Digital Signal Processing,
Prentice-Hall, 1975.
[3] Sneha N. Kherde, Meghana Hasamnis, "Efficient Design and
implementation of FFT", International Journal of Engineering Science and
Technology (IJEST), ISSN : 0975- 5462, NCICT Special Issue Feb. 2011
[4] K. R. Rao and P. Yip, Discrete Cosine Transform: Algorithms
advantages and Applications, Academic Press, New York, 1990
[5] R M Goudar and Priya Pise Compression Technique using DCT and
Fractal Compression - A Survey Advances in Physics Theories and
Applications ISSN 2224-719X(Paper) ISSN 2225-0638 Vol 3, 2012.

11