International Journal of Computer Applications (0975 – 8887) Volume 179 – No.12, January 2018 1 Validation of Image Compression Algorithms using Neural Network Nikhilesh Joshi Research Scholar, Dept of Computer Engg Thadomal Shahani Engineering College, Bandra(W), Mumbai Tanuja K. Sarode, PhD Professor & Head, Dept of Computer Engg, Thadomal Shahani Engineering College, Bandra(W), Mumbai ABSTRACT We live in Digital Era where information is generated at rapid space. Images constitute a major part of information. It becomes essential to use image compression techniques in order to reduce storage space and transmission bandwidth. Image compression algorithm can be validated using Neural Network. In this paper various methods of Image compression such as BTC, DCT, DWT are optimized and Validate using neural network. This is achieved by comparing methods based on set of parameters. . The resultant compression metrics are calculated and visual quality of image is analyzed. Neural network implementation is done based on two different methods desired matrix and entropy based method. Experimental analysis shows 60 % reduction in storage space requirement and effective optimization using different methodology. General Terms Image Processing, Neural Networks Keywords Image Compression, Entropy, Block Truncation Coding, Discrete Cosine Transform(DCT) Discrete wavelet tranform(DWT, Image Quality Metrics, Neural Network, Back propogation. 1. INTRODUCTION Now a day‟s most of the information is in the form of Images. Images require large amount of space for storage and consumes more bandwidth during transmission. Image Compression plays a vital role in reducing the storage space requirement and helps to increase transmission ratio over network. A gray scale image that is 256 x 256 will have 65, 536 pixels to store and a typical 640 x 480 color image have nearly a million. Downloading of these files from internet can be very time consuming task. Image data comprise of a significant portion of the multimedia data and occupy the major portion of the communication bandwidth for multimedia transmission. The image compression technique most often used is transform coding. Transform coding is an image compression technique that first switches to the frequency domain, then does its compression. The transform coefficients should be decor related, to reduce redundancy and to have a maximum amount of information stored in the smallest space [1] [2]. Two fundamental components of compression are redundancy and irrelevancy. Redundancies reduction aims at removing duplicate information from the signal source (image/video). Irrelevancy reduction omits parts of the signal that will not be noticed by the signal receiver In digital image compression, three basic data redundancies can be identified and exploited: 1. Coding redundancy 2. Inter pixel redundancy 3. Psycho visual redundancy Data compression is achieved when one or more of these redundancies are reduced or eliminated. Coding redundancy use shorter code words for the more common gray levels and longer code words for the less common gray levels. This is called Variable Length Coding. To reduce this redundancy from an image we go for the Huffman technique were we are assigning fewer bits to the more probable gray levels than to the less probable ones achieves data compression. Inter pixel redundancy is directly related to the inter pixel correlations within an image. Because the value of any given pixel can be reasonable predicted from the value of its neighbors, the information carried by individual pixels is relatively small. Much of the visual contribution of a single pixel to an image is redundant; it could have been guessed on the basis of its neighbor‟s values Psycho visual redundancy: Human perception of the information in an image normally does not involve quantitative analysis of every pixel or luminance value in the image. In general, an observer searches for distinguishing features such as edges or textural regions and mentally combines them into recognizable groupings. The brain then correlates these groupings with prior knowledge in order to complete the image interpretation process. Thus eye does not respond with equal sensitivity to all visual information. Certain information simply has less relative importance than other information in normal visual processing. This information is said to be psycho visually redundant. The elimination of psycho visually redundant data results in a loss of quantitative information. The Organization of this paper is as follows. The Performance Measure are described in Section II, DCTBTC in section III, Experimental Analysis of DCTBTC in section IV, DWTDCT and its experimental analysis in section V, Neural network design and its different method, experimental result in section VI, Comparative study in section VII and Conclusion in Section VIII 2. PERFORMANCE PARAMETER Image Quality measure plays important role in various image processing applications. Once an image compression technique is designed and implemented its performance evaluation has to be done. The performance metrics needs to be considered in such a way to be able to compare results
8
Embed
Validation of Image Compression Algorithms using Neural ... · becomes essential to use image compression techniques in order to reduce storage space and transmission bandwidth. Image
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
International Journal of Computer Applications (0975 – 8887)
Volume 179 – No.12, January 2018
1
Validation of Image Compression Algorithms using
Neural Network
Nikhilesh Joshi Research Scholar, Dept of Computer Engg Thadomal Shahani Engineering College,
Bandra(W), Mumbai
Tanuja K. Sarode, PhD Professor & Head, Dept of Computer Engg,
codeword are assigned to the corresponding symbols according
to the probability of the symbols. The entropy encoders are used
to compress the data by replacing symbol represented by the
equal length codes with the code word‟s whose length is
proportional to corresponding probability
8. CONCLUSION An enhanced BTC algorithm was proposed getting better
image quality after compression. The method uses low pass
filtering on the image first and then takes the DCT of given
image, then implement the BTC. By applying DCT after low
pass filtering the number of pixels required for compression is
reduced. A set of standard images were tested using different
block size and image quality metrics were calculated. It was
found that the reconstructed image has better quality
compared to BTC. For 4x4 Block size 25% compression was
achieved. Experimental results show that the difference
between the pixel values of the original and the reconstructed
image is considerably reduced. The test results also show the
performance of the proposed method based on the parameters
PSNR, MSE, Correlation and CR. The results show that the
PSNR and MSE values are high when compared with BTC,
1
4096
2
1
2
80
1
9
Original
Image
256 x 256
pixels
Reduced
Image
64 x64
pixels
O
U
T
P
U
T
International Journal of Computer Applications (0975 – 8887)
Volume 179 – No.12, January 2018
8
even when the compression ratio is same. It gives a better
enhancement in the visual quality of the reconstructed images
even at the edges. Further time taken by DCTBTC for
encoding is less when compared with BTC as it involves
simple calculations. The research works attempt to
systematically design and then validate the compression
technique. After the validation scope for optimization in that
area of application will be taken in to consideration.
Regarding DWTDCT It can be concluded that result obtain
indicate reduction in storage space requirement. This will
directly help in reducing the transmission bandwidth
requirement for various images. The second method provides
better result compared to the first one. Moreover it is observed
that time required to perform compression is reduce compared
to traditional approach. In future these methods can be
enhanced further by applying some other standard algorithm
in mixed mode.
It can be concluded that result obtain indicate reduction in
storage space requirement. This will directly help in reducing
the transmission bandwidth requirement for various images.
The design of NN helps in reducing the time requirement for
calculating the compression ratio. The Neural network also
helps in validating the result obtained before training by
comparing them to result obtained after training
Table 8 Comparative study of NN methodologies Image Desired
Matrix
Value
Neural
Network
Entropy
Based
Difference
Img1 62 70 70 0
Img2 53 60 60 0
Img3 58 60 70 10
Img4 58 70 70 0
Img5 60 60 70 10
Img6 65 70 70 0
Img7 65 80 70 10
Img8 66 70 70 0
Img9 71 60 60 0
Img10 71 60 60 0
Img11 73 60 60 0
Img12 71 60 60 0
Img13 78 50 70 20
Img14 68 70 70 0
Img15 69 60 60 0
Img16 50 60 70 10
Img17 66 70 70 0
Img18 73 60 70 10
Img19 55 50 70 20
Img20 69 70 60 10
9. ACKNOWLEDGMENT I would express my gratitude to my guide Dr Tanuja Sarode
for her constant support and guidance. I would like to thank
Prof Arun Kulkarni, Dean (AICTE Affairs) Dept of IT,
Thadomal Shahani Engineering College for constant support,
appreciation and advice in right direction. I am also thankful
to Dr. G T Thampi, Principal TSEC, Dr. Subhash. K Shinde,
Chairman Board of Studies Mumbai University for the
valuable inputs and corrections suggested during course of
research. I would also like to thank my wife Sonali and son
Shlok for giving me space and time to carry out my work.
10. REFERENCES [1] Rafael C. Gonalez, Richard E.Woods “Digital Image
Compression” 3rd Edition, Prentice Hall.
[2] E. J.Delp, O R. Mitchelle “ Image Compression using Block truncation Coding” IEEE Transaction Communication 27(9) (1979) 1335-1342
[3] M. D. Lema, O.R . Mitchelle “Absolute Moment Block ytruncation Coding and its Application to Color Images” IEEE Transcaction Communication Vol COM-32, No.10, pp1148-1157 Oct 1984
[5] U.Y Desai, M.M. Muzuki, B.K.P.Horn “Edge and mean based Compression” MIT Artifical Intelligence Laboratory AI Memo No.1584,November 1996.
[6] T.M.Ammarunnishad, V.K Govindan, T.M Abraham ” Improving BTC Image Compression using a Fuzzy Complement Edge operator”Signal Processing Letters, Vol 88. Issue 12 December 2008 pp. 2989-2997
[7] T.M.Ammarunnishad, V.K Govindan, T.M Abraham “ A Fuzzy Complement edge operator “ IEEE proceeding of the Fourteen International Conference on Advance Computing and Communication Mangalore, Karnataka, India December 2006
[8] Aditya Kumar, Pradeep Singh “ Futuristic Algorithm for Gray Scale Image Based on Enhanced Block Truncation Coding” International Journal of Computer Information system Vol 2 No.5 pp 53-60 ,2011
[9] Jaymol Mathews, Madhu S Nair, Liza Jo ” Modified BTC Algorithm for Gray Scale Images Using max-min Quantizer” IEEE Transaction 2013 pp 377-382
[10] Manish Gupta, Dr. Anil Kumar Garg “Analysis of Image Compression Algorithm Using DCT”, International Journal of Engineering Research and Applications ISSN:2248-9662, Vol 2 Issue 1 Jan-Feb2012 pp 515-521
[11] Mahinderpal Singh, Meenakshi Garg “Mixed DWT-DCT Approached Based Image Compression Technique” International Journal of Engineering and Computer Science ISSN:2319-7242, Vol 3 Issue 11 November 2014 pp 9107-9111
[12] Bhavna Sagwan, Mukesh Sharma, Krishan Gupta “RGB based KMB Image Compression Technique” International Conference on Reliability, Optimization and Information Technology Feb 2014
[14] Mahinderpal Singh, Meenakshi Garg “Mixed DWT-DCT Approached Based Image Compression Technique” International Journal of Engineering and Computer Science ISSN:2319-7242, Vol 3 Issue 11 November 2014 pp 9107-9111
[15] Bhavna Sagwan, Mukesh Sharma, Krishan Gupta “RGB based KMB Image Compression Technique” International Conference on Reliability, Optimization and Information Technology Feb 2014
[16] M.Singh and M Garg, “Mixed DWT DCT Approached
Based Image Compression technique ” International
Journal of Engineering and Computer Science ISSN
2319-7242 volume 3 Issue 11 Nov 2014page 9008 -9111.