Image compression using sand algorithm

INTERNATIONAL Computer Engineering and Technology ENGINEERING –
International Journal of JOURNAL OF COMPUTER (IJCET), ISSN 0976 &
6367(Print), ISSN 0976 – 6375(Online) Volume 3, Issue 3, October-December (2012), © IAEME
TECHNOLOGY (IJCET)
ISSN 0976 – 6367(Print)
ISSN 0976 – 6375(Online)
Volume 3, Issue 3, October - December (2012), pp. 121-128
IJCET
© IAEME: www.iaeme.com/ijcet.asp
Journal Impact Factor (2012): 3.9580 (Calculated by GISI) ©IAEME
www.jifactor.com

IMAGE COMPRESSION USING SAND ALGORITHM

Karimella Vikram1, Dr. Niraj Upadhayaya2, Dr.A.Goverdhan3, T.Madhavi Kumari4
1
Associate Professor and HOD, Dept of CSE, Medak College of Engineering and Technology,
Kondapak, Medak Dist
Research Scholar, CSE Dept., Jawaharlal Nehru Technological University, hyderabad-515002,
Andhra Pradesh, India
Email: vikramkariella@yahoo.com
2
Professor, Dept. of CSE,
JBIET Moinabad , hyderbad
Andhra Pradesh, India
Email: : drniraju@gmail.com
3
Professor, Dept. of CSE
JNTUH
4
Professor, Dept. of CSE
JNTUH

ABSTRACT

In today’s world transfer of information plays a vital role. Today the term information is not just
confined to literal data. Images and videos convey a lot more than text messages. But
transmission and storage of images have their own setbacks. The major one being the space and
time complexity and this is where the compression of images comes into picture. Ours is one
such algorithm indigenously developed by me, which has the potential to compress the images
with out any loss of any information. This makes the compression algorithm ideal for domains
where loss of information is unacceptable. One major application can be in the transfer of
information on stellar objects by non-terrestrial telescopes, spacecrafts and space probes, which
sends back vital information and it is also necessary to ensure the bandwidth and also the cost of
transmission is low. It can also be employed in bio-telemetry and other related fields where
information transmitted has to be reliable (without any loss) for faithful analysis of the data and
at the same time the cost factor should also be considered. We have an ideal solution for all these
domains. We haven’t confined our self to the theoretical proof of this algorithm; rather we have
also come up with a practical working model of the algorithm which implements a section of the
algorithm to show its practical feasibility. The software developed works on the basis of Sand

121

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976 –

algorithm; it is demo software, which can be implemented on 8 bit (256 color .bmp) color
images.

I. INTRODUCTION

AND works on the principle of elimination of repetitions, it literally absorbs pixels of the same
nature (color) as the name suggests .In other words the pixels of the same nature converges to a
single point this reduces the number of pixels which in turn reduces the size of the image there
by increasing the rate of transmission. The coding of the information is so done that during the
decompressing process not a single bit of information is lost.

II. ILLUSTRATION
Any digitized image can be broken into pixels and the number of pixels per area is proportional
to the resolution of the image (fig 0). These pixels hold information for recreating the image. If
the picture is a 16 bit image then each pixel would hold 16 bits of information as to what color
has to be displayed. For example if you were to take a photograph of yourself in a studio with a
red screen as a background then the region except your image in the photograph would be red .It
is this region that the sand takes advantage of, it sucks all the pixel into almost a single pixel.
The absorption process has to take place in such a way that it can be recreated with out any loss.
Thus the compression is a sequential process the flow diagram of which is shown in fig.

A. Algorithm

122


The flow diagram shown above gives an idea of the various processes involved in the
compression process each of which will be discussed in the following sections. As it is known,
that every picture is made up of thousands of pixels and each pixel has its colour and these pixels
in combination form the image. That would mean each pixel requires 32 bits to represent a pixel.
The picture below shows an illustrative view of the arrangement of pixels in a picture.

Fig-(a): Zoomed View of the Pixel Arrangement

The compression process can be broadly classified into two sections where one process follows
the other and they are termed as Latitudinal compression Longitudinal compression To
understand the compression process in a better way let us take a segment of the picture and
implement the compression process on that. The block in fig (1) shows a section of a picture with
pixels arranged, each circle represents a pixel

Fig-(1) Latitudinal Compression

The block shown in fig (1) represents a block that would undergo latitudinal compression. In the
latitudinal compression process the block is first divided into rows (5 rows in this example), each
row is processed separately. The processing involves taking the first pixel and comparing them
with the consecutive pixel. Depending on the result of the comparison, further processing is
carried is carried out. If the color of first pixel does not match with the second pixel, both the
pixels are retained. If both pixels tend to have the same color then the first pixel is retained while
the second pixel is eliminated or absorbed. This process is applied to each row of the picture.
Employing the same logic on the fig (1) we get fig(1.1). Taking the first row we have two blue
pixels that are absorbed into one, likewise in the fourth row, three red pixels are absorbed into
one pixel (absorbed regions are represented by dots). Similar process is carried out through out.
This process is applied to each row of the picture. Employing the same logic on the fig (1) we get
fig(1.1). Taking the first row we have two blue pixels that are absorbed into one, likewise in the
fourth row, three red pixels are absorbed into one pixel (absorbed regions are represented by
dots). Similar process is carried out through out. The fig(1.1) shows the pixel block after the
application of latitudinal compression.

123


Latitudinal Compression

The fig(1.1) Single bit absorption table

After the latitudinal compression, the next step involves the creation of “Single bit absorption
table“(SBA table). It is this table, which gives information regarding the absorption pattern of the
pixels. A simple logic is followed for this, Absorbed pixels in the block are represented as “0”
and the unabsorbed pixels are represented as”1”. Thus we get the table as seen in fig . The size of
the pixel is proportional to the size of the block.

Single Bit Absorption Table (for Latitudinal Comp

Fig

2) Longitudinal Compression
This is similar to the latitudinal compression the only difference being that the pixel block after
the latitudinal compression is taken and is divided into columns and the compression process is
carried out vertically (column wise) columns and the compression process is carried out
vertically (column wise)

This block is obtained from fig-2.1 by discarding the dots (absorbed pixels)

Transmission / storage box

Fig Fig

124


3) Efficiency of Compression
The efficiency of the compression depends completely on the picture and the arrangement of the
pixels. The efficiency varies from picture to picture. Considering the blocks taken here the
efficiency would be given as follows. If the picture taken here is a 24 bit color picture then the
memory occupied by the pixel block before the compression process would be
Siz_before = (N * W)
Where,
N = Number of pixels in the block.
W= Weight of each pixel.
Thus,
Siz_before = (25 * 24) =600
After the compression process,
Siz_after = (Nr * W) + OH
Where,
Nr = Number of pixels remaining after
Compression
OH = Over head due to the SBA tables
OH = (2 * N) = (2 * 25) = 50.
Siz_after = (13 * 24) + 50 = 362

Thus, this yields a reduction of nearly 40% of the original size of the block. Practically higher
compression ratio is realizable considering the number of pixels in the pixel block to be much
higher.

B. Retrieving your Image
So far we have been talking about compressing the image. More important than this is retrieving
the image without any loss .As said earlier the information at the receiving end would be the
resultant block (fig 3.1) and the two tables (fig 3.2,3.3) .We need to create the Image using these
three information units. The decompressing algorithm works conversely to that of the
compression algorithm. It is responsible for decoding the information from the received
“Transmission box”. It makes use of a simple concept; the processing takes place as follows. The
two tables are analyzed bit by bit; if the bit in both the tables is “1” then the pixel is got from the
“Resultant block”. This process is depicted in the diagram shown below, where the shaded
(yellow) region represents the bit being processed. As seen since both pixels are “1” the pixel
from the resultant block is moved to the “Recreation block” The second condition is that when
the bit in both the SBA table is “0” then the pixel gets the color of the pixel on the left in the
recreation block. The example for which is shown below The third condition is encountered
when the bit in the first SBA table is “1” and the second table has the value as “0”. In such a
case the pixel gets the color of the pixel above it in the “Recreation table” The example for
which is show below. Note that the “Resultant block” comes into picture only when both the
table have the data as “1”. For all other conditions the “Recreation box” itself acts as the source
for the pixels. The illustration for the third condition is shown below. The processing is carried
out in the similar fashion; every bit The two tables are analyzed bit by bit; if the bit in both the
tables is “1” then the pixel is got from the “Resultant block”. This process is depicted in the
diagram shown below, where the shaded (yellow) region represents the bit being processed. As
seen since both pixels are “1” the pixel from the resultant block is moved to the “Recreation

125


block” The second condition is that when the bit in both the SBA table is “0” then the pixel gets
the color of the pixel on the left in the recreation block. The example for which is shown below
The third condition is encountered when the bit in the first SBA table is “1” and the second table
has the value as “0”. In such a case the pixel gets the color of the pixel above it in the
“Recreation table” The example for which is show below. Note that the “Resultant block” comes
into picture only when both the table have the data as “1”. For all other conditions the
“Recreation box” itself acts as the source for the pixels. The illustration for the third condition is
shown below. The processing is carried out in the similar fashion; every bit

The second condition is that when the bit in both the SBA table is “0” then the pixel gets the
color of the pixel on the left in the recreation block. The example for which is shown below

The third condition is encountered when the bit in the first SBA table is “1” and the second table
has the value as “0”. In such a case the pixel gets the color of the pixel above it in the
“Recreation table” The example for which is show below. Note that the “Resultant block” comes
into picture only when both the table have the data as “1”. For all other conditions the
“Recreation box” itself acts as the source for the pixels. The illustration for the third condition is
shown below.

The processing is carried out in the similar fashion; every bit in the SBA table is interpreted
based on the decompression rules described earlier. Thus ultimately the transmitted image can be
regenerated as follows
Recreated Image

C. Compression Criteria
After every compression process the algorithm checks if the criteria is satisfied which as given
below
(Siz_after + OH) < Siz_before only If this criteria is satisfied the compression process carried
out.
D. Exponential Reduction:
The compression process can be further be aggravated by a process of exponential reduction
where your “Resultant block” can be further reduced by the same compression process until the

126


sand criteria is broken. Not only the pixel can be compressed but also the table can be reduced
the table which is in the binary format can be grouped into “W” bits (24 bits) to form a colored
pixel which can be rearranged and then can be compressed to reduce the total over head. Further
more like the longitudinal and latitudinal compression cross compression is also possible. The
type of compression to be adopted depends the arrangement of the pixels and the efficiency it
yields.

III. CONCLUSION

The abstract presented here gives a over view of the compression idea developed by me. The
Sand algorithm is a promising idea waiting to be tapped practically. The algorithm has extensive
scope in various fields of research and analysis. Though this algorithm cannot match the
efficiency of other lossy compression formats like (JPEG) .It would be an ideal choice in areas
where quality of information cannot be compensated for the cost. This algorithm will enable us
to cut cost and increase the rate of transmission of images with out any adverse effect on the
quality of the image.

REFERENCES

[1] IEEE Transaction on Image Compression entitled “New image compression techniques using
multiwavelets and multiwavelet packets”, Apr 2001, Volume: 10 Issue:4, PP 500 - 510
[2] IEEE Conference Proceedings on July 01-July 03 2008 entitled “A new compression
technique for binary text images”
[3] Data and Image Compression: Tools and Techniques, 4th Edition by Gilbert Held (Author),
Thomas R. Marshall (Author)

AUTHORS

K.Vikram is working as a Vice-Principal at Medak College of Engineering
and Technology (MCET), Hyderabad, India. He has received MCA and
ME.(Computer Science and Engineering) from Anna University. Presently,
he is a Research Scholar. He has published and presented good number of
technical papers in National and International Conferences. His main
research interests are Software Engineering, Data Mining, and Network
Security.

Dr. Niraj Upadyay is working as a Dean of CSE Dept. at JBIET. He has
received B.Tech and MS.(Computer Science and Engineering) from BITS and
Ph.D. . He has published and presented good number of technical papers in
National and International Conferences. His main research interests are
Software Engineering, Data Mining, and Ne

127


Dr. Govardhan has been conferred CSI Chapter Patron Award (2010- 2011),
Shining Image of India Award 2011, Seva Chakra Puraskar 2011, Bharat Jyoti
Award 2010, Rajiv Gandhi Excellence Award 2010, Best Citizens of India
2010 Award, Life Time
Achievement Gold Medal Award 2010, Rashtriya Vidya Saraswati Puraskar
2010, Eminent Educationist Award 2010 and Certificates of Excellence for
Outstanding services, Achievements,
Contributions, Meritorious Services, Outstanding Performance and Remarkable Role in the field
of Education and Service to the Nation He has been listed as One among the • Top Three Faculty
in JNTU Hyderabad, made by Outlook Survey for the year 2008
• He was awarded Telugu Vignana Parithoshikam, by the Government of Andhra Pradesh for
B.E(CSE) program
• He was nominated as the Best Student for the year 1987- 88 in A.P. Residential Junior College
(APRJC), Nagarjuna Sagar, A.P.
• He was awarded Merit Scholarship by the Government of Andhra Pradesh for securing first
position in School. He was awarded PADMA JYOTHI for securing First Position in School in
10th standard.
• He is a Member in Executive Council, JNTUH, He is a member of Standing Committee for
Academic Senate, JNT University Hyderabad and Academic Advisory Committee (AAC), UGC-
Academic Staff College and Sports Council, JNT University Hyderabad
• He is the Chairman for Post Graduate Board of Studies (BOS) in Computer Applications, Yogi
Vemana University, Kadapa, Member on various BOS ( for CSE and IT, UG and PG), including
JNT University Hyderabad, Adikavi Nannaya University, Rajahmundry, Satavahana University,
Karimnagar
and VR Siddhartha Engineering College, Vijayawada and Sree Nidhi Institute of Science and
Technology, Hyderabad.
• He is the Vice-Chairman for CSI Hyderabad Chapter

• He is a Member on the Editorial Boards for Eight International .Mrs.Thoomati
Madhavi Kumari obtained B.Tech. in Electronics & Communication
Engineering and M.Tech. in Computer Science from JNT University. Presently
Mrs. Madhavi is pursuing Ph.D. in Data security for natural languages using
FPGA. Mrs. Madhavi joined the faculty of ECE Department of JNU College of
Engineering, Kukatpally, Hyderabad as Assistant Professor and served the
department for 13 years. Later she was appointed as Assistant Director,
UGCASC, JNT University, Hyderabad Mrs. Madhavi, in the ECE Department,
conducted two faculty development programs in collaboration with IBM and Intel. In the
department, she was associated with the implementation of AICTE sponsored project
on computer networks. Mrs. Madhavi guided many M.Tech. and MCA

128

Image compression using sand algorithm

Recommended

More Related Content

What's hot (17)

Viewers also liked (20)

Similar to Image compression using sand algorithm (20)

More from IAEME Publication (20)

Image compression using sand algorithm