Image compression technique is used in many applications for example, satellite imaging, medical imaging, video where the size of the iamge requires more space to store, in such application image compression effectively can be used. There are two types in image compression techniques Lossy and Lossless comression. Both these techniques are used for compression of images, but these techniques are not fast. The image compression techniques both lossy and lossless image compression techniques are not fast, they take more time for compression and decompression. For fast and efficient image compression a parallel computing technique is used in matlab. Matlab is used in this project for parallel computing of images. In this paper we will discuss Regular image compression technique, three alternatives of parallel computing using matlab, comparison of image compression with and without parallel computing.

1. Journal for Research| Volume 01| Issue 09 | November 2015
ISSN: 2395-7549
All rights reserved by www.journalforresearch.org 6
Fast and Efficient Image Compression based on
Parallel Computing using MatLab
Prakash Venkateshappa
M. Tech Student Associate Professor
Department of Electronics & Communication Engineering Department of Electronics & Communication Engineering
MSEC MSEC
Abstract
Image compression technique is used in many applications for example, satellite imaging, medical imaging, video where the size
of the iamge requires more space to store, in such application image compression effectively can be used. There are two types in
image compression techniques Lossy and Lossless comression. Both these techniques are used for compression of images, but
these techniques are not fast. The image compression techniques both lossy and lossless image compression techniques are not
fast, they take more time for compression and decompression. For fast and efficient image compression a parallel computing
technique is used in matlab. Matlab is used in this project for parallel computing of images. In this paper we will discuss Regular
image compression technique, three alternatives of parallel computing using matlab, comparison of image compression with and
without parallel computing.
Keywords: Lossy and Lossless compression, DWT, IDWT, Huffman coding, parallel computing, PCT
_______________________________________________________________________________________________________
I. INTRODUCTION
Image compression technique is used in many applications for example, satellite imaging, medical imaging, video where the size
of the iamge requires more space to store, in such application image compression effectively can be used. There are two types in
image compression techniques Lossy and Lossless comression.
 Lossy image compression
 Lossless image compression
1) Lossy image compression
In lossy data compression original data is not exactly restored after decompression and accuracy of re-construction is traded with
efficiency of compression. Lossy data compression algorithms are transform coding (for example, discrete cosine tranform),
Karhunen-Loeve Transform (KLT) and wavelet based coding(for example, continuous wavelet transform-CWT and Discrete
wavelet transform(DWT).
2) Lossless Image Compression
As the name implies, lossless image compression schemes exploit redundancies without incurring any loss of data. Thus, the data
stream prior to encoding and after decoding is exactly the same and no distortion in the reconstruction quality is observed.
Lossless image compression is therefore exactly reversible.
II. BLOCK DIAGRAM OF IMAGE COMPRESSION SYSTEM
The figure 1 describes the image compression system; this block diagram explains the Encoding and Decoding of image
compression technique with Huffman coding technique.
Fig. 1: Block diagram of Image compression system

2. Fast and Efficient Image Compression based on Parallel Computing using MatLab
(J4R/ Volume 01 / Issue 09 / 002)
All rights reserved by www.journalforresearch.org 7
Discrete Wavelet Transform (DWT)A.
The DWT can be taken at as the multi resolution decomposition of a sequence. It takes a sequence z (n) of length N and produces
ayield of length N. The output can be seen and analyzed as the multi resolution representation of the sequence z (n), and has N/2
qualities at the highest resolution and N/4 values at the next resolution and etc. That is the frequency resolution is low at the
higher frequencies and high at the lower frequencies, while the time resolution is high at the higher frequencies and low at the
lower frequencies.
It principally comprises of multiplying the input sequence by translates and dilates of the wavelet DWT (Discrete wavelet
Transform) module comprising of a HPF(High pass filter) and LPF(low pass filter) followed by down sampling unit to compute
the approximation and detailed coefficients of the input. The input sequence X (n) is utilizing high pass and low pass filters to
produce the output Yd(n) and Ya(n) representing the approximation and detail samples of the input signal X (n).
Inverse Discrete Wavelet Transform (IDWT)B.
The IDWT (Inverse Discrete Wavelet Transform) can be dissected as the multi resolution decomposition of a sequence. It takes a
Ya(n),Yb(n) as input and up sample these input, and then convolute them with HPF and LPF to produce the output. The output
can be seen as the multi resolution representation of X (n).
Quantization EncodingC.
Assume I is an M × N gray image, its pixels can be described as follows.
I = {xij| 1 ≤ i ≤ M, 1 ≤ j ≤ N, xij  {0, 1, 2… 255} } (1)
After performing DWT on I, we obtain four sub-bands LL, HL, LH, and HH of size M/2× N /2. Conventionally, JPEG2000
performs the uniform quantization on the resulting DWT. Coefficients sub-bands. For most images, after subtracting the average
of maximum and minimum, the distribution of coefficients is similar to a zero-mean Laplacian and hence uniform quantization
adopts value 0 as the center of quantization [8].
Quantization DecodingD.
The de quantization formula is shown below.
Rq(i, j) =
{
(Q(i, j −) + r)∆b + ω, Q(i, j) > 0
(Q(i, j −) + r)∆b + ω, Q(i, j) < 0
0, others
Where Rq(i,j) and Q(i, j) stand for the reconstructed coefficient and the quantized value 0 r 1 is an optional parameter for
controlling the recovering position within quantization interval. Finally, perform the IDWT (Inverse Discrete Wavelet
Transform) on the recovered coefficients in all 4 sub-bands to obtain the reconstructed image I [8].
Huffman Coding StepsE.
First1)
1) With decreasing probability sort the gray levels.
2) Two smallest probabilities have to add.
3) Then the new values into the list have to sort.
4) Then repeat 1st
and 3rd
step until only two probabilities remains.
Second2)
1) For the highest probability give the code 0, and code 1 for the lowest probability in the summed pair.
2) Then have go backwards through the tree one node and repeat from 1 until all gray levels have a unique code.
III.PARALLEL COMPUTING USING MATLAB(PCT)
Parallel computing has been considered to be "the high end of computing", and has been used to model difficult scientific and
engineering problems found in the real world. Parallel computing can be used in image compression for fast result of
compression and decompression process. In this project Parallel computing Toolbox (PCT) is used for parallel computing in
Matlab. There are three alternatives are there for parallel computing in Matlab, those are bcMPI, Star-P, Parallel computing
toolbox (PCT).
BcMPIA.
bcMPI is an open source software library that is an alternative to MatlabMPI and is geared towards large, shared supercomputer
centers. The bcMPI library was developed at the Ohio Supercomputer Center (OSC) to provide an efficient, scalable
communication mechanism for parallel computing in MATLAB while maintaining compatibility with the MatlabMPI API
(Hudak et al., 2007) [7].

3. Fast and Efficient Image Compression based on Parallel Computing using MatLab
(J4R/ Volume 01 / Issue 09 / 002)
All rights reserved by www.journalforresearch.org 8
Star-PB.
Star-P is a client-server parallel computing platform for MATLAB available from Interactive Supercomputing. Star-P supports
fine grained parallel as well as embarrassingly parallel modes of operation. The biggest advantage offered by Star-P is that it
eliminates the need for the developer to use explicit Message Passing Interface (MPI) message passing calls for communicating
between the back-end processes. By using the “*p” construct, users can simply indicate the variables or data that are meant to be
distributed over the back-end processes [7].
Parallel Computing ToolboxC.
The Parallel Computing Toolbox (PCT) along with the MATLAB Distributed Computing Server (MDCS) are commercial
products offered by The MathWorks Inc. The PCT provides functions for parallel for-loop execution, creation/manipulation of
distributed arrays as well as message passing functions for implementing fine grained parallel algorithms.The MATLAB
Distributed Computing Server (MDCS) gives the ability to scale parallel algorithms to larger cluster sizes. The MDCS consists
of the MATLAB Worker processes that run on a cluster and is responsible for parallel code execution and process control. The
PCT also allows users to run up to 8 MATLAB Labs or Workers on a single machine. This enables interactive development and
debugging of parallel code from the desktop. After parallel code has been developed, it can be scaled up to much larger number
of Worker or Labs in conjunction with the MDCS [7].
IV.EXPERIMENTAL RESULT
In this project fast and efficient image compression technique based on parallel computing is proposed and developed using
MatLab Parallel Computing Toolbox. A set of images are taken to experiment the effectiveness of the algorithm. The figure 2
sows the various images used in the experiment. The experimental result with proposed compression method has been arranged
in the Table1. Table 1 show that compression and de compression without parallel computing is not fast as compared to our
proposed algorithm.
Fig. 2: Original images used in this project
Table - 1
Experimental Result
Image Width Height
Compression
Ratio
Encoding
without parallel
(seconds)
Decoding
without parallel
(seconds)
Parallel
computing
CR
Encoding
With parallel
computing
(seconds)
Decoding
With parallel
computing
(seconds)
1 512 467 54.122925 29.487857 32.500330 55.368042 3.502272 4.958386
2 512 467 54.457092 28.641213 32.273664 55.926514 1.978338 4.829220
3 450 356 54.280853 28.640774 32.334073 59.237671 2.002130 4.448639
4 307 230 43.379211 28.769231 33.677322 46.842957 1.980165 5.909391
5 1419 1001 44.053650 28.744912 33.408402 45.030975 1.962430 6.285207
V. CONCLUSION
The design of fast and efficient image compression using parallel computing is presented in this work. Matlab coding is used to
design the proposed architecture. Then the design is simulated on Matlab. The proposed architecture is able to perform fast

4. Fast and Efficient Image Compression based on Parallel Computing using MatLab
(J4R/ Volume 01 / Issue 09 / 002)
All rights reserved by www.journalforresearch.org 9
compression of images with acceptable compression ratio. Finally, the proposed system is very fast and efficient in compression
of all types of images such as nature images, medical images, satellite images etc.
REFERENCES
[1] B.C Vemuri, S.Sahni, F.chen, C.Leonard, and J.Fitzsimmons, “Lossless Image compression”, Dept of CISE, Univ. of Florida, Gainsvelle,F1.32611
[2] Ronald A. Devore, Bjorn Jawerth, and Bradley J. Lucier, “Image compression Through wavelet Transform Coding”, IEEE Transactions on information
theory,Vol.38,No.2 March 1992.
[3] Dipalee Gupta, Siddhartha Choubey, “Discrete Wavlet Transform for Image Pocessing, International Journal of Emerging Technology and Advanced
Engineering, ISSN 2250-2459, Volume 4, Issue 3, March 201.5.
[4] Jagdish H. Pujar, Lohit M. Kadlaskar, “A NEW LOSSLESS METHOD OF IMAGE COMPRESSION AND DECOMPRESSION USING HUFFMAN
CODING TECHNIQUESFPGA”, Journal of Theoretical and Applied Information Technology.
[5] Diya Chudasama, Khushboo Parmar, Dipali Patel, Kruti J. Dangarwala, Shaishav Shah, “Survey of Image Compression Method Lossless Approach”,
International Journal of Engineering Research & Technology (IJERT), ISSN: 2278-0181, Vol. 4 Issue 03, March-2015.
[6] Tejas S. Patel, Ravindra Modi, Keyur J. Patel, “Image Compression Using DWT and Vector Quantization”, International Journal of Innovative Research in
Computer and Communication Engineering,Vol. 1, Issue 3, May 2013
[7] Ashok Krishnamurthy, Siddharth Samsi and Vijay Gadepally, “Parallel MatLab Techniques”, Ohio Supercomputer Center and Ohio State University
U.S.A.
[8] Pro-Yueh Chen and Jia-Yu Chang, “An Adaptive Quantization Scheme for 2-D DWT Coefficients”, International Journal of Applied Science and
Engineering 2013.
[9] Rajesh K. Yadav, S.P. Gangwar and Harsh V. Singh Preetam Bhosle and Hari Krishna Moorthy, “Study and analysis of wavelet based image compression
techniques”, International Journal of Engineering, Science and Technology, Vol. 4, No. 1, 2012, pp. 1-7, 2012
[10] M. Mozammel Hoque Chowdhury and Amina Khatun, “Image Compression Using Discrete Wavelet Transform”, IJCSI International Journal of Computer
Science Issues, Vol. 9, Issue 4, No 1, July 2012
[11] Ramandeep Kaur Grewal M.Tech (I.T.), Navneet randhawa Neenu Joseph, “IMAGE COMPRESSION USING DISCRETE COSINE TRANSFORM &
DISCRETE WAVELET TRANSFORM”, International Journal of Computing & Business Research ISSN (Online): 2229-6166,2012.
[12] Priyadarshini K S, Dr. G S Sharvani, “A Survey on Parallel Computing of Image Compression Algorithms”,International journal of innovative technology
and research, ISSN 2320-5547, April 2015.
[13] Bhavesh Vekariya1, Chhaya Patel, “A Review on Image Compression in Parallel using CUDA”, International Journal of Engineering Research and
Development e-ISSN: 2278-067X, p-ISSN: 2278-800X, www.ijerd.com Volume 11, Issue 02 February 2015.
[14] Arnold Meijster, “Efficient Sequential and Parallel Algorithms for Image Processing”, Second International Conference on Computer and Electrical
Engineering,IEEE, pp:765-771777,2011.