A Presentation of Dissertation Phase-III on “Fast, Efficient, Lossless Image Compression System” Presented By Akshay Gajanan Bhosale Under the Guidance of Prof. S. N. Kore Walchand College of Engineering, Sangli
Welcome message from author
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
A Presentation
of Dissertation Phase-III on
“Fast, Efficient, Lossless Image Compression System”
Presented By Akshay Gajanan Bhosale
Under the Guidance of Prof. S. N. Kore
Walchand College of Engineering, Sangli
• Brief Introduction
• Work Done
• Modifications
• Results
• Future work
2
Contents
Prediction Template
3
• There are 4 cases for this prediction template.
Main Flowchart
4
Encode first two pixels directly
Start
Get next pixel P, decide H and L from two reference pixels
Apply Golomb-Rice code
Is current pixel P L ≤ P ≤ H?
Apply Adjusted Binary code
Stop
Yes
No
Distribution Model
• This is the distribution model used in FELICS algorithm
• It is used to decide whether to go for adjusted binary code or to use Golomb-Rice code 5
• Developed a fast, efficient & lossless image compression system
• Used the FELICS algorithm to compress different types of images
• Collected results for different images
6
Work Done
• 4 reference pixels are taken instead of 2• Average of reference pixels is calculated • Deviation of current pixel is taken from average and it is
encoded
7Intensity of pixels
Probability
Modification I
N2 N3 N4
N1 P1
Reference Template
Modification II• In this modification, the algorithm is not lossless, instead it
combination of lossy and lossless techniques• We used a lossy 2D-DWT technique and lossless FELICS
algorithm
8
Image 2D-DWT FELICSAlgorithm
CompressedImage
Block Diagram
Results• For different types of images we get different Compression ratios (
CR)
• The quality of original and recovered images are compared using different performance metrics• Mean Square Error (MSE)• Peak Signal to Noise Ratio (PSNR)• Structural Content (SC)• Normalized Cross-Correlation (NCC)• Normalized Absolute Error (NAE)• Maximum Difference (MD)• Average Difference (AD)
• Summary9
• Implement CDP algorithm for colour images
10
Further Improvements
PAPERS:• [1] Tsung-Han Tsai, Yu-Hsuan Lee, and Yu-Yu Lee, “Design and Analysis of High-
Throughput Lossless Image Compression Engine using VLSI-Oriented FELICS Algorithm” IEEE trans. on VLSI Systems, Vol. 18, No.1, January 2010.
• [2] D. Huffman, “A method for the construction of minimum redundancy codes,” Proc. IRE, vol. 140, pp. 1098–1011, Sep. 1952.
• [3] T.Welsh, “A technique for high-performance data compression,” IEEE Comput., vol. 17, no. 6, pp. 8–10, Jun. 1984.
• [4] P. G. Howard and J. S. Vitter, “Fast and efficient lossless image compression,” in Proc. IEEE Int. Conf. Data Compression, 1993, pp. 501–510.
• [5] X. Wu and N. D. Memon, “Context-based, adaptive, lossless image coding,” IEEE Trans. Commun., vol. 45, no. 4, pp. 437–444, Apr. 1997.
• [6] M. J. Weinberger, G. Seroussi, and G. Sapiro, “The LOCO-I lossless image compression algorithm: Principles and standardization into JPEG-LS,” IEEE Trans. Image Process., vol. 9, no. 8, pp. 1309–1324, Aug. 2000.
• [7] P.-C. Tseng, Y.-C. Chang, Y.-W. Huang, H.-C. Fang, C.-T. Huang, and L.-G. Chen, “Advances in hardware architectures for image and video coding-a survey,” Proc. IEEE, vol. 93, no. 1, pp. 184–197, Jan. 2005.
11
References
References
WEB RESOURCE:• ieeexplore.ieee.org• en.wikipedia.org
BOOKS:
• Rafael Gonzalez, Richard E. Woods, “Digital Image Processing”, Pearson Education, India (2002).
12
ANY QUESTIONS ?
THANK YOU
Test Images used
14
back
Compression ratio (CR)• Compression ratio is given by
CR =
• Higher the value of compression ratio, more the is image compressed
• Higher compression ratio degrades the quality of image
15
Results for Compression Ratio
Test ImagesMethod
Lena Baboon Boat Bridge Satellite Image
Medical Image
FELICS 1.72 1.26 1.55 1.37 0.99 2.32
JPEG 11.13 5.31 8.67 5.81 3.53 18.19
DWT+FELICS 5.33 3.86 5.07 4.22 3.64 7.07
16
back
Mean Square Error (MSE)• Mean Square Error is defined as,
MSE =
• The large value of MSE means that image quality is poor
17
back
Test ImagesMethod
Lena Baboon Boat Bridge Satellite Image
Medical Image
FELICS 0 49.22 0 0 0.2144 1.445
JPEG 17.16 101.30 29.07 72.23 302.79 2.82
DWT+FELICS 45.637 314.004 90.5728 175.95 883.191 50.566
Peak Signal to Noise Ratio• Peak Signal to Noise Ratio (PSNR) is defined as,
PSNR =
• Small value of PSNR means the image quality is poor
18
back
Test ImagesMethod
Lena Baboon Boat Bridge Satellite Image
Medical Image
FELICS 99 31.2092 99 99 54.81 46.53
JPEG 35.43 27.18 33.49 29.54 23.31 43.62
DWT+FELICS 31.53 23.16 28.56 25.67 18.67 31.09
Structural Content (SC)• Structural Content is defined as,
SC =
• The large value of Structural Content (SC) means that image is of poor quality
19
back
Test ImagesMethod
Lena Baboon Boat Bridge Satellite Image
Medical Image
FELICS 1 0.9995 1 1 0.9999 0.9993
JPEG 1.00035 1.00014 1.0005 1.0014 0.9644 1.00082
DWT+FELICS 0.99733 1.01104 0.999621 1.00915 1.0542 1.0055
Normalized Cross Correlation• Normalized Cross-Correlation (NCC) is defined as,
NCC =
• Value of NCC close to 1, means the image quality is good
20
back
Test ImagesMethod
Lena Baboon Boat Bridge Satellite Image
Medical Image
FELICS 1 0.9988 1 1 1.00001 1.00031
JPEG 0.9993 0.9971 0.9989 0.9969 1.0078 0.9995
DWT+FELICS 1.00005 0.9860 0.9978 0.9899 0.9445 0.9965
Normalized Absolute Error• Normalized Absolute Error (NAE) is defined as,
NAE =
• Value of NAE close to 0, means the image quality is good
21
back
Test ImagesMethod
Lena Baboon Boat Bridge Satellite Image
Medical Image
FELICS 0 0.0414 0 0 0.000475 0.00440
JPEG 0.0240 0.05811 0.0310 0.0547 0.1147 0.0064
DWT+FELICS 0.03110 0.0944 0.0443 0.0786 0.1844 0.0279
Maximum Difference (MD)• Maximum Difference (MD) is defined as,
MD =
• Value of MD close to 0, means the image quality is good
22
back
Test ImagesMethod
Lena Baboon Boat Bridge Satellite Image
Medical Image
FELICS 0 40 0 0 25 9
JPEG 42 63 46 72 143 15
DWT+FELICS 82 111 129 97 200 50
Average Difference (AD)• Average Difference is defined as,
AD =
• The large value of AD means that image quality is poor
23
back
Test ImagesMethod
Lena Baboon Boat Bridge Satellite Image
Medical Image
FELICS 0 0.00176 0 0 -0.00583 -0.05930
JPEG 0.0044 -0.0113 0.0097 0.0085 -0.1502 0.0685
DWT+FELICS -0.37313 -0.36531 -0.3748 -0.1387 1.0568 0.6572