A VIDEO COMPRESSION TECHNIQUE UTILIZING SPATIO-TEMPORAL LOWER COEFFICIENTS

http://www.iaeme.com/IJECET.asp 10 editor@iaeme.com
International Journal of Electronics and Communication Engineering & Technology
(IJECET)
Volume 7, Issue 1, Jan-Feb 2016, pp. 10-19, Article ID: IJECET_07_01_002
Available online at
http://www.iaeme.com/IJECETissues.asp?JType=IJECET&VType=7&IType=1
Journal Impact Factor (2016): 8.2691 (Calculated by GISI) www.jifactor.com
ISSN Print: 0976-6464 and ISSN Online: 0976-6472
© IAEME Publication
A VIDEO COMPRESSION TECHNIQUE
UTILIZING SPATIO-TEMPORAL LOWER
COEFFICIENTS
Ashwini Atulkar
Department of Electronics and Communication Engineering,
SATI Engineering College, Vidisha, M.P
Abhishek Jain
Asst. Professor, Department of Electronics & Communication Engineering
SATI Engineering College, Vidisha, M.P
ABSTRACT
With the advancement of communication in recent trends, video
compression plays an important role in the transmission of information on
social networking and for storage with limited memory capacity. Also the
inadequate bandwidth for transmission and lower quality make video
compression a serious phenomenon to consider in the field of communication.
There is a need to improve the video compression process which can encode
the video data with low computational complexity with better quality along
with maintaining speed. In this work, a new technique is developed based on
the block processing utilizing the lower coefficients between frames. The
performance parameters such as Compression Ratio (CR) & Peak Signal to
Noise Ratio (PSNR) are evaluated based on the algorithm for several standard
video sequences. The proposed method results in higher compression rates as
compared to the spatial block based compression technique. The simulation
result shows that the proposed algorithm sustain faithful compression and
reproduction of the video, preserving the picture quality.
Key words: Video Compression, Block Processing, Spatial.
Cite this Article: AshwiniAtulkar and Abhishek Jain. A Video Compression
Technique Utilizing Spatio-Temporal Lower Coefficients. International
Journal of Electronics and Communication Engineering & Technology, 7(1),
2016, pp. 10-19.
http://www.iaeme.com/IJECET/issues.asp?JType=IJECET&VType=7&IType=1

A Video Compression Technique Utilizing Spatio-Temporal Lower Coefficients
1. INTRODUCTION
The high requirement of bandwidth has been the major problem in a video. A typical
system needs to send several frames per unit second so that the illusion of a moving
picture can be generated. For instance, a colour video clip recorded with a resolution
of 1920x1080 assuming 30 frames per second requires memory of 1.5 GB for second.
This means even a DVD of capacity 5 GB cannot handle a video of 4 minutes
duration. It is exactly for this reason video compression has become very significant.
To eliminate the redundancy, each individual frame is coded. Also with motion
compensation, a great amount of redundancy can be eliminated between two
successive frames. In spite of the continuous development in the storage space and
transmission bandwidth now-a-days, as the amount of data obtained by digitizing the
video signal is huge, compression is very important in most of the digital video
applications.
A video is nothing but a sequence of images which are displayed one at a time.
The continuous image sequence bluffs us that something is moving. There are very
small changes in the video frames, which make one frame differ from the other. So all
this makes one requirement of video compression as storage and transfer get easier.
The compression of video tends two type of technique first is lossy compression
where compression of video frame is done in a way that the original information is
irreversible. Although this technique makes lots of space in disks and bandwidth, but
due to loss of information it is undesirable.
The other method is lossless in which original information can be successfully
recovered after applying decompression technique. This paper has also developed
similar kind of lossless technique.
In compression method those information which is getting repeated in every frame
is removed from the further frame [1, 2]. This identification of frame position and
storage of that information is done in the video compression method. This is also
categorized in inter and intra frame information compression where consecutive
frames are used for finding similar data. While in case of intra-frame fault tolerance
capacity increases as if one of the consecutive frames lost then video can be built-up
with other frames available in the compressed one. Video is categorized in two type
first is black and white while another is color. In case of black and white, gray format
is used while in the case of color RGB color format is used in the frame. Now
compression technique in the video is classified on the basis of spatial as well as on
temporal way. Here in the case of spatial, pixel value of a frame is identified by the
pixel value of the same frame. While in the case of the temporal, pixel value of a
frame is identified by the pixel value of the other co-related frame.
2. RELATED WORK
Liyin et.al [3] presented the Full Search method, which is the most computationally
and more costly for all algorithm of block matching. By this algorithm, at each
possible location in the search window calculates the cost function and gives the
highest PSNR amongst any block matching algorithm and it finds the best possible
match. To attain the same PSNR doing as little computation as possible by Fast block
matching algorithms. Variable Size Block Matching Methods Attaining, good video
compression ratio and image quality are a conflicting requirement, specially with
fixed size block matching (FSBM), where the size of all the blocks is the same. The
above problem can be overcome by variable size block matching motion estimation
technique (VSBME). In variable-size block matching (VSBM) [4]- [5] smaller blocks

Ashwini Atulkar and Abhishek Jain
can be used to depict complex motion while larger blocks can be used in areas where
the image content is stationary or undergoing uniform motion while the no of blocks
and bit rate are fixed. Quad tree-Structured Variable-Size Block-Matching Motion
Estimation [6][7] presents two algorithms with Minimal Error. The first algorithm
computes the optimal selection of variable-sized blocks to provide the best achievable
prediction error under the fixed number of blocks for a Quad tree-based VSBM
technique. Although this algorithm is computationally intensive, it does provide a
yardstick by which the performance of other more practical VSBM techniques can be
measured. The video signal is an integral part of multimedia which has a tremendous
importance in most of the applications involving the concept of the multimedia i.e.
video conferencing, broadcast digital video, and high- definition television (HDTV),
etc.[8]. Mueller introduces the generalized Gaussian distribution to model the DCT-
coefficients more accurate than with Laplace distributions [9]. Mathematical
morphology can be considered as a shape oriented approach to signal processing and
some of its features make it very useful for compression for videos [10]. Man et.al
proposed a novel Four-Step Search Algorithm for Fast Block Motion Estimation. The
proposed algorithm is based on the center-based global minimum motion vector
distribution characteristic of real world image sequence, a new Four-Step Search
algorithm for fast block-based motion estimation [11]. A fast block- matching
algorithm that uses three fast matching error measures, besides the conventional
mean-absolute error (MAE) or mean-square error (MSE) is provided in [12]. Shen
and Edward [13] proposed video compression using Based Rate Scalable Method; it is
a new wavelet based rate scalable video compression algorithm. This technique is
referred as the Scalable Adaptive Motion Compensated Wavelet (SAMCoW)
algorithm. Edmund and Joseph proposed the distribution of DCT-coefficients in the
field of image compression and an approximation of the AC-coefficients with Laplace
distributions [14]. H.R. Kusuma discussed a comparison of various spatial and
temporal redundancy removal techniques used for video compression[15]. The
performance of Discrete Cosine Transform (DCT) based algorithm for compressing
video with perceiving quality and content of video presented in [16]. M. Atheeshwari
et.al.[17] presented compression techniques necessary for video processing especially
to find how much amount of data to compressed.
3. PROPOSED WORK
The proposed work is based on basic block processing spatial technique as well as
Inter-frame video compression technique. In this methodology, the video compression
is done by using a 16×16 DCT block processing with a 6 coefficient mask by
detecting and eliminating the similarities between the consecutive frames. The
schematic diagram of the proposed methodology is shown in Fig.3 & Fig.4 and the
major blocks are as follows:
Video Reader
Video is a collection of images which is called as a frame. Here these collections of
images are displayed in so fast sequence that human eyes could not judge that they
actually see one image at a time. As the contents of the consecutive frames are mostly
same, but change in object position is new information on the frame. So a reading of
video means conversion of video in a sequence of frames of RGB format.

Consecutive Frames
Here frames are passed in pair of two, where the pairing of frames is done on the basis
of their sequence. This sequence of frames is consecutive. So two consecutive frames
are passed for each run. This can be understood as let V be video then this contain
frames as {f1, f2, f3, ……………fn}. So a pair of consecutive frames is (f1, f2), (f2,
f3), (f3, f4), etc.
Figure 1 Collection of video frame
16x16 DCT Block Processing with 6 coefficient Mask
In this step 16 ×16 DCT block processing is done, which simply means that the
frames of the video is divided into blocks of 16 ×16 and then apply DCT on each
block. Now the mask of 6×6 coefficient is applied to each block. By multiplying the
DCT matrix by mask the elements of the matrix become zero, due to this the
information get reduced. This is shown in Fig. 2.
IDCT Processing
In this step simple inverse process of DCT is done. Here each 16x16 block is
recombine to make single frame.
Euclidean distance
Now the difference between the consecutive frames is evaluated in this step. This
distance helps in identifying the frame similarity. As difference pixel is zero means
those pixel position is unchanged in both frames. This can be understood as Let X be
a frame and Y be the other consecutive frame from same video then distance can be
calculated as:
Sparse Matrix
Here Non zero position in the distance matrix need to be saved in a matrix for future
reference, when video need to be read. So sparse matrix is created for the frame which
store non matching pixel value of the frames
Frames

Single frame
16x16 blocks
6x6 Masking
1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0
1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Figure 2 16x16 block DCT application
Here one check is applied which helps in identifying that either the new matrix
size is more or less as compared to the previous one. If the size of a matrix is less,
than it is stored in sparse matrix otherwise save it as an original frame matrix. As it is
already known that sparse matrix contain three columns first represent row, second
represent column and third represent value at that position.

Figure 3 Block diagram of proposed compression model
Combine Frame
In this step, all matrices of the frame combine in one cell and use store in the disk.
This combination of matrix is of different size as different frames may have their
different size of the matrix.
Reconstruct Frame
In case of de-compression, reverse process is required where each matrix of frames is
reconstructed with the help of previous frame information. This can be understood as
if frame is constructed from the sparse matrix, then it provides pixel value of the
information stored in sparse matrix, then rest of pixel value is obtained from the
previous frame.
Figure 4 Block diagram of proposed De-Compression model
Proposed Compression Algorithm
Input: V // V: Input Video
Output: CV // CV: Compressed Video
1. VRead_Video(V)
2. FF[n]Consecutive_Frames(V) // FF[n]: n consecutive frame
3. Loop 1:n
4. BBlocks(FF[n])
5. BMask_DCT(B)
Combine Frames
Sparse Matrix
Euclidean distance
IDCT Block Processing
16x16 DCT Block Processing
with 16 Coefficient Mask
Consecutive Frames
Read Video
Combine Frame
Reconstructed Frames
Consecutive Frames
Read Video

6. FF[n]IDCT(B)
7. DEuclidean_dist(FF[n]) // D is distance matrix
8. Loop 1:x // x, y are row and column of video frame
9. Loop 1:y
10. If D[x, y]) > 0
11. SM[m][x, y, D[x,y]) // m number of positions
12. EndIf
13. EndLoop
14. EndLoop
15. EndLoop
4. EXPERIMENT AND RESULT
The proposed methodology is implemented for video compression and the results are
compared with spatial method. Experimental results validate the effectiveness of the
approach. All algorithms and utility measures were implemented using the MATLAB
tool. The tests were performed on a 2.27 GHz Intel Core i3 machine, equipped with 4
GB of RAM, and running under Windows 7 Professional.
Evaluation Parameters
Peak Signal to Noise Ratio (PSNR)
It is used to find the amount of data present from the received signal as it may corrupt
by the presence of some noise. So it is termed as the peak signal to noise ratio. The
ratio between the maximum possible data and the noise that affects the fidelity of its
representation.
Compression Ratio (CR)
In this parameter storage size of the original video is compare with the video obtain
after applying compression technique.In other word, the compression ratio is the ratio
between uncompressed video size and compressed video size.
Where, C.R. = Compression Ratio
Data Sets
An experiment done on the standard video name crocus, snowflake, etc. as well as on
the artificial videos, Results technique:
(a) (b)

(c) (d)
Figure 5 Original Video
(a) (b)
(c) (d)
Figure 6 Compressed Video
Table 1 Average value of PSNR for different standard videos
PSNR Comparison
Video Spatial Proposed Work
Crocus 48.62 62.46
Snowflake 43.34 68.89
Fp 47.53 55.42
Star 54.12 56.23
Table.2 Average value of Compression Ratio for different standard videos
Compression Ratio
Video Spatial Proposed Work
Crocus 1.20 1.21
Snowflake 1.28 1.29
Fp 1.42 1.43
Star 1.42 1.43
It has been observed by Table 1 & 2 that video compression by proposed work is
better as compared to previous one, as PSNR value is higher. It is observed that as the
number of the frames increases, then there is a chance of getting less PSNR value.
5. CONCLUSIONS
Large number of video compression technique developed for different requirement.
Proposed method utilizes the basic block processing combined with temporal lower
coefficient compression method. It has been obtained that the proposed work is better
as compared to the spatial method on the basis of different performance parameters. It
is also obtained that video size affects the parameters like compression ratio, PSNR. If
size of a video is small better compression ratio is obtained. As the size of a video is
increased, compression ratio gets decreased. As there are various possibilities still

present in the work like compression of grey video, inter-frame information
compression etc.
REFERENCES
[1] T.A. Browne, J.V. Condell, G. Prasad, T.M. McGinnity, An Investigation into
Optical Flow Computation on FPGA Hardware, International Machine Vision
and Image Processing Conference, 2008, 176-181.
[2] FransciscoBarranco, MatteoTomasi, JavierDiaz, Mauricio Vanegas and Eduardo
Ros Parallel Architecture for Hierarchical optical Flow Estimation based On
FPGA, IEEE Transactions On very large integrated VLSI System, Volume 20,
Issue 6, June 2012, 1058-1067.
[3] XieLiyin, Su Xiuqin, Zhang Shun, A Review of Motion Estimation Algorithms
for Video Compression, International Conference on Computer Application and
System Modeling (ICCASM 2010), 2010.
[4] H. Jelveh and A. Nandi, Improved variable size block matching motion
compensation for video conferencing applications, in Digital Signal Processing,
A. Cappellini and A. Constantinides, Eds. Berlin, Germany: Springer-Verlag,
1991.
[5] G.R. Martin, R.A. Packwood and I. Rhee, Variable Size Block Matching Motion
Estimation with Minimal Error, Department of Computer Science, University of
Warwick, Oct95.
[6] I. Rhee, G. R. Martin, S. Muthukrishnan and R. A. Packwood, Quad tree-
Structured Variable-Size Block- Matching Motion Estimation with Minimal
Error, IEEE Trans. on Circuits and Systems for Video Technology, Vol. 10,pp.
42-50, Feb. 2000.
[7] Xiao-lin Chen, Design of UAV Video Compression System Based on H.264
Encoding Algorithm, International Conference on Electronic & Mechanical
Engineering and Information Technolo-gy.Vol-5, pp.2619-2622, 2011.
[8] C. Pandey, S. Kumar and R. Tiwari, An Innovative Approach towards the Video
Compression Methodology of the H.264 Codec: using SPIHT Algorithms,
International Journal of Soft Computing and Engineering (IJSCE) ISSN: 2231-
2307, vol.2, Issue-5, November 2012.
[9] F. Mueller, Distribution Shape of Two-Dimensional DCT Coefficients of Natural
Images, Electronic Letters, vol. 29, pp. 1935–1936, 1993.
[10] P. Salembier, P.Brigger, J.R.Casas and M.Pardas, Morphological Operators for
Image and Video Compression, IEEE Transactions on Image Processing, vol. 5,
June 1996.
[11] Lai-Man Po and Wing-Chung Ma, A Novel Four-Step Search Algorithm for Fast
Block Motion Estimation, IEEE Trans. Circuits and Systems for Video
Technology, vol. 6, June 1996.
[12] Y.C. Lin and S.C. Tai, Fast Full-Search Block-Matching Algorithm for Motion-
Compensated Video Compression, IEEE, vol. 45, 1997.
[13] K.Shen and E.J. Delp, Wavelet Based Rate Scalable Video Compression, IEEE
transactions on circuits and systems for video technology, vol. 9, 1999.
[14] R. Reininger and J. Gibson, Distributions of the Two Dimensional DCT
Coefficients for Images, IEEE Transactions Communication, vol. 31, pp. 835–
839, June 1983.
[15] H. R. Kusuma and M. rao, Video Compression Using Spatial and Temporal
Redundancy-A Comparative Study, International Journal of Innovative Research
in Science, Engineering and Technology, vol. 4,pp. 4802-4808.

[16] S. S. Wadd, B. S. Patil, Video Compression Using DCT, International Journal of
Advanced Research in Computer Science and Software Engineering, Volume 4,
Issue 9, September 2014.
[17] M.Atheeshwari, K.Mahesh, Video Compression Techniques – A Comprehensive
Survey, International Journal of Advanced Research in Computer Science and
Software Engineering, Volume 4, Issue 1, January 2014.
[18] B.K.N.Srinivasa Rao, P.Sowmya. Architectural Implementation of Video
Compression through Wavelet Transform Coding and EZW Coding.
International Journal of Electronics and Communication Engineering &
Technology, 3(3), 2012, pp. 202 - 210.
[19] Dr. Uma B.V, Dr. Geetha K.S, Dr. Prasanna Kumar S.C and Naveen Kumar L.
Implementation of H.264 Based Real Time Video Compression Algorithm for
Underwater Channel. International Journal of Electronics and Communication
Engineering & Technology, 5(4), 2014, pp. 43 - 49.

A VIDEO COMPRESSION TECHNIQUE UTILIZING SPATIO-TEMPORAL LOWER COEFFICIENTS

More Related Content

What's hot (15)

Viewers also liked (20)

Similar to A VIDEO COMPRESSION TECHNIQUE UTILIZING SPATIO-TEMPORAL LOWER COEFFICIENTS (20)

More from IAEME Publication (20)

Recently uploaded (20)

A VIDEO COMPRESSION TECHNIQUE UTILIZING SPATIO-TEMPORAL LOWER COEFFICIENTS