JOURNAL OF TELECOMMUNICATIONS, VOLUME 16, ISSUE 1, SEPTEMBER 2012

Color Image Compression Using Two Dimensions Discrete Wavelet Transform (2D-DWT)
Samir J. AL- Muraab
ABSTRACT: More applications of advanced communication systems (Mobile, Internet, Satellite) need small or narrow channel capacity in order to increase the information bit rate/sec, so we use the information Compression process. This study presents a system for color image compression type JPEG by using Two Dimension Discrete Wavelet Transform [2D- DWT], with some digital signal processing techniques. The results of this system are 50% compression with good accuracy and 75% compression with suitable accuracy, and with consideration the image size for display process. All programs, graphics, and the results are accomplished by using MATLAB version (7). Index Terms Two Dimensions Discrete Wavelet Transform 2D-DWT, Color Image Compression, Color Image Decompression.

Samir J. AL- Muraab is with the Babylon University- college of Engineering.

1. INTRODUCTION
For huge volume of multimedia (audio, video and graphics) and without compression we need more storage capacity and large transmission bandwidth. The increasing use of digital images and growth of data intensive multimedia based web applications have cause much pressure on the researchers to find the way of using the images in the web applications more effectively. Internet teleconferencing, High Definition Television (HDTV), satellite communications and digital storage of movies are not feasible without a high degree of compression. As it is, such applications are far from realizing their full potential largely due to the limitations of common image compression techniques [1 2 4]. Two main common compressed graphic image formats are namely Joint Photographic Experts Group (JPEG) which is established by International Standards Organization and International Electro-Technical Commission while the Graphic Interchange Format (GIF) method is commonly used for line art and other images in which geometric shapes in the Internet[1 2 3 4].

3. WAVELETS FOR IMAGE COMPRESSION

By contractions and translations of mother wavelet on the input data the wavelet transform uses both the spatial and frequency correlation of data. It supports the multi-resolution analysis of data i.e. it can be applied to different scales according to the details required, which allows progressive transmission and zooming of the image without the need of extra storage. And because of symmetric nature that is both the forward and the inverse transform have the same complexity, building fast compression and decompression routines of wavelet transform. So that the characteristics of wavelet transform well corrected for image compression include the ability to take into account of Human Visual Systems (HVS) characteristics, very good energy compaction capabilities, robustness under transmission, high compression ratio etc.[2]. The sub band coding scheme is very similar to the scheme of implementation of wavelet compression: the signal is decomposed using filter banks. The output of the filter banks is down-sampled, quantized, and encoded. The decoder decodes the coded representation, up-samples and recomposes the signal [2]. We can get from information of an image, approximation and detail sub signals by using wavelet transform. The approximation sub signal shows the general trend of pixel values and other three detail sub signals show the vertical, horizontal and diagonal details or changes in the images. The details can be set to zero without significantly changing the image if it very small (less than threshold). The greater the compression ratio means the greater the number of zeros. If the energy retained (amount of information retained by an image after compression and decompression) is 100% then the compression is lossless as the image can be reconstructed exactly. This occurs when the threshold value is set to zero, meaning that the details have not been changed. If any value is changed then energy will be lost and thus lossy compression occurs [2 11 12].

2. IMAGE COMPRESSION
Two elementary components of compression are redundancy and irrelevancy reduction. The removing of duplication from the signal source image means redundancy reduction. And the parts of the signal that is not noticed by the signal receiver means irrelevancy reduction, namely the Human Visual System (HVS). In general, three types of redundancy can be identified [4]: (i) Spatial Redundancy or correlation between neighboring pixel values, (ii) Spectral Redundancy or correlation between different color planes or spectral bands and (iii) Temporal Redundancy or correlation between adjacent frames in a sequence of images especially in video applications. Removing the spatial and spectral redundancies as much as possible is mean that reducing the number of bits needed to represent an image which is the goal of compression.

JOURNAL OF TELECOMMUNICATIONS, VOLUME 16, ISSUE 1, SEPTEMBER 2012

4.

JPEG COLOR IMAGE COMPRESSION METHOD USING HAAR WAVELET TECHNIQUE

In this paper, the color images are considered instead of gray-scale image. The wavelet transforms and compression steps and processes are equally same to the gray-scale image but for the color images have three color components of the images, so that when we are converted the pixel values in an image from RGB colors to YIQ colors we must computed the wavelet transform for each components of the image separately. These mean that the information of Y, then I, and finally Q. Also means that the color image divided into three layers. Then we are used the scaling function() equations and the box basis of Haar wavelets(!) equations as shown in equations (1) &(2) below [2 8].

After that we are calculated the compression ratio (CR) which it is meaning as ( the number of nonzero elements in original matrix t/the number of the nonzero elements in updated transformed matrix) for the three different types of thresholding as in eq.(3)above, then calculated also the PSNR and the Man Squared Error (MSE) which is depended on original and approximation of decompressed images as equations (4)&(5) shown below.

(1)

Fig.1:Normal JPEG Color Image

(2)

The PSNR for color scale image (8 bits/pixel) is defined by-

(4)
Three types of thresholding are applied, namely, hard thresholding, soft thresholding, and universal thresholding [2 8]. These types of thresholding are stated as shown in equation (3) below[2 8]:-

Where MSE is the Man Squared Error defined by-

(5)
Where I is original image, I1 is the approximation of decompressed image, and m &n are dimensions of the image.

(3)

The PSNR values & compression ratio for different threshold values and techniques are shown in table (1). Fig. (2-9) show the original, transmitted, and the decomposition JPEG color images for various values of and thresholding. When increasing the value of we show that the calculations results of PSNR in case the soft thresholding are less than the hard thresholding and also the compression ratio. These mean that the performance of hard thresholding are the best CR. And the performances of the soft thresholding are better CR than the universal methods. So that the characteristics of each one of the three thresholding types (CR, PSNR, and image quality and resolution) are accepted depending on the type of applications.

5.

RESULTS & DISCUSSION

We are taken the 88 color image from a specific portion of a typical image shown in Fig.1. This image is represented as matrix (2D array), then all the operation of averaging and differencing to get a new matrix to complete the process.

JOURNAL OF TELECOMMUNICATIONS, VOLUME 16, ISSUE 1, SEPTEMBER 2012

10

Table1: Compression Ratio & PSNR Values for Various and Thresholding

Original Image Fig. 1 Fig. 1 Fig. 1 Fig. 1 Fig. 1 Fig. 1 Fig. 1

Transformed Image Fig 2 Fig 4 Fig 6 Fig. 8 Fig. 10 Fig. 12 Fig. 14

Threshold No threshold Hard, =10 Hard, =80 Hard, =180 Soft, =80 Soft, =180 Universal threshold

Compression Ratio, PSNR CR=44%, PSNR=14.14 db CR=44.4%, PSNR=14.1 db CR=68%,PSNR=14.07 db CR=74% ,PSNR=13.767 db CR=68%,PSNR= 11.625 CR=73.7%, PSNR= 7.74 CR= 59%, PSNR=14

Decompressed Image Fig. 3 Fig. 5 Fig. 7 Fig. 9 Fig. 11 Fig. 13 Fig. 15

Fig.2:Transmitted JPEG Image (No Thresholding)

Fig.3: Decomp. JPEG Image (No Thresholding)

Fig.4: Transmitted JPEG Image (Hard Thresholding =10)

Fig.5: Decomp. JPEG Image (Hard Thresholding =10)

Fig.6:Transmitted JPEG Image (Hard Thresholding =80)

Fig.7: Decomp. JPEG Image (Hard Thresholding =80)

Fig.8:Transmitted JPEG Image (Hard Thresholding =180)

Fig.9: Decomp. JPEG Image (Hard Thresholding =180)

Fig.10:Transmitted JPEG Image (Soft Thresholding =80)

JOURNAL OF TELECOMMUNICATIONS, VOLUME 16, ISSUE 1, SEPTEMBER 2012

11

Fig.11: Decomp. JPEG Image (Soft Thresholding =80)

Fig.12:Transmitted JPEG Image (Soft Thresholding =180)

Fig.13: Decomp. JPEG Image (Soft Thresholding =180)

Fig.14:Transmitted JPEG Image (Universal Thresholding)

Fig.15: Decomp. JPEG Image (Universal Thresholding)

REFERENCES

[10] Gersho, A. and Gray, R. M., Vector Quantization and

Signal Compression, Kluwer Academic Publishers, 1991. [11] Nelson, M., The Data Compression Book, 2nd edition., M&T books,November.1995, [11] Robert L. Cook and Tony DeRose, Wavelet Noise, ACM Transactions on Graphics, July 2005, Volume 24 Number 3, Proc. Of ACM SIGGRAPH 2005, pp. 803811. [12] G. Beylkin, R. Coifman, and V. Rokhlin, Fast wavelet transforms and numerical algorithms, I. Communications on Pure and Applied Mathematics, 44(2): 141-183, March 1991. [13] Colm Mulcahy, Image compression using the Haar Wavelet transforms, Internal Report. [14] H. Marmolin, Subjective MSE measures, IEEE Trans. Systems Man. Cybernet. 16, 1986, 486489. [15] James D. Foley, Andries van Dam, et. al.Computer Graphics : Principles and Practice, Addison-Welsey, Reading, MA, 2nd Edition, 1990.
Samir Jasam Mohammad (Member IEEE) was born in Babylon-1959, Iraq. He received the B.Sc. degree in Electronics and Communications Department from the University of Baghdad (1984)-Iraq, M.Sc. and Ph.D. degrees in Electronics and Communication Engineering from the University of Technology-Iraq in 1987 and 2004 respectively. Since 2004, he has been with the University of Babylon-Iraq, where he is lecturer in Electrical Engineering Department. His research interests include DVB, CDMA, Modulation Technique, Image processing.

[1] Talukder, K.H. and Harada, K., A Scheme of Wavelet

Based Compression of 2D Image, Proc. IMECS, Hong Kong, pp. 531-536, June 2006. [2] Talukder, K.H. and Harada, K., Haar Wavelet Based Approach for Image Compression and Quality Assessment of Compressed Image, JAENG Intern-ational Journal of Applied Mathematics,36:1,IJAM_36_1_9,1 February 2007. [3] Ahmed, N., Natarajan, T., and Rao, K. R., Discrete Cosine Transform, IEEE Trans. Computers, vol. C-23, Jan. 1974, pp. 90-93. [4] Pennebaker, W. B. and Mitchell, J. L. JPEG, Still Image Data Compression Standards, Van Nostrand Reinhold, 1993. [5] Rao, K. R. and Yip, P., Discrete Cosine Transforms Algorithms, Advantages, Applications, Academic Press, 1990. [6] Wallace, G. K., The JPEG Still Picture Compression Standard, Comm. ACM, vol. 34, no. 4, April 1991, pp. 30-44. [7] Eric J. Stollnitz, Tony D. Derose and David H. Salesin, Wavelets for Computer Graphics- Theory and Applications Book, Morgan Kaufmann Publishers, Inc. San Francisco, California. [8] Vetterli, M. and Kovacevic, J., Wavelets and Subband Coding, Englewood Cliffs, NJ, Prentice Hall, 1995, . [9] ISO/IEC/JTC1/SC29/WG1 N390R, JPEG 2000 Image Coding System, March,1997.