Empirical evaluation of EZW and other encoding techniques in the wavelet-based image compression domainby A. Suruliandi, S. P. Raja

Int. J. Wavelets Multiresolut Inf. Process.


Applied Mathematics / Information Systems / Signal Processing


2nd Reading

April 17, 2015 13:51 WSPC/S0219-6913 181-IJWMIP 1550012

International Journal of Wavelets, Multiresolution and Information Processing

Vol. 13, No. 2 (2015) 1550012 (20 pages) c© World Scientific Publishing Company

DOI: 10.1142/S0219691315500125

Empirical evaluation of EZW and other encoding techniques in the wavelet-based image compression domain

A. Suruliandi and S. P. Raja∗

Department of CSE, Manonmaniam Sundaranar University

Tirunelveli, Tamilnadu, India ∗avemariaraja@gmail.com

Received 25 March 2014

Revised 3 March 2015

Accepted 6 March 2015

Published 21 April 2015

This paper discusses about embedded zerotree wavelet (EZW) and other wavelet-based encoding techniques employed in lossy image compression. The objective of this paper is two fold. Primarily wavelet-based encoding techniques such as EZW, set partitioning in hierarchical trees (SPIHT), wavelet difference reduction (WDR), adaptively scanned wavelet difference reduction (ASWDR), set partitioned embedded block (SPECK), compression with reversible embedded wavelet (CREW) and space frequency quantization (SFQ) are implemented and their performance is analyzed. Second, wavelet-based compression schemes such as Haar, Daubechies and Biorthogonal are used to evaluate the performance of encoding techniques. The performance parameters such as peak signalto-noise ratio (PSNR) and mean square error (MSE) are used for evaluation purpose.

From the results it is observed that the performance of SPIHT encoding technique is providing better results when compared to other encoding schemes.

Keywords: Wavelet image compression; EZW; SPIHT; WDR; ASWDR; SPECK; SFQ;

CREW. 1. Introduction

Digital images have been widely used in numerous application areas. Massive digital images have been recently applied in forensic science, medical imaging, remote sensing, security purposes, etc. Once personal computers gained the capacity to display sophisticated pictures as digital images, people started to seek methods for efficient representation of these digital pictures in order to simplify their transmission and save disk space. So the need for image compression is started in this point. The field of image compression has a wide spectrum ranging from classical lossless techniques and popular transform approaches to the more recent segmentation-based (or second generation) coding methods.

Image compression techniques can be classified into lossless and lossy techniques.12 With lossless compression, the original image is recovered exactly after 1550012-1 2nd Reading

April 17, 2015 13:51 WSPC/S0219-6913 181-IJWMIP 1550012

A. Suruliandi & S. P. Raja decompression. Much higher compression ratios were obtained and it is usually difficult to find errors between the decompressed image and the original image.

Lossy compression has few data losses in the decompressed image. In many cases, it is not possible that there could be error-free reproduction of the original image.

In such a case, the limited amount of error introduced by lossy compression may be acceptable.

JPEG31 for still images and the MPEG standard15 for video images are based on DCT.20 However, some image compression algorithms are based on DWT.3,5,13

Some of the compression algorithms are used in the earlier days based on the wavelet transform introduced by Davis6 and Mallat.14 In addition to the above, many such innovative techniques were developed by Alkholidi2 and Xiang.32 Also, compression of color images has attracted many researchers.1,7 1.1. Image compression theory

The usual steps involved in compressing an image are wavelet decomposition quantization and encoding. 1.1.1. Wavelet decomposition

Wavelet is a basic function that is isolated with respect to time or spatial location.

A wavelet transform can be utilized to decompose a signal into component wavelets.

After this is done the coefficients of the wavelets can be decimated to remove some of the details. Wavelets have the great advantage of being able to separate the fine details in a signal.

Figure 1 shows the general steps involved in wavelet-based image compression. Processing may involve compression, encoding, denoising and quantization.

Wavelets also used in speech compression. They are used in biomedical applications, edge detection, feature extraction, etc.

Wavelets find application in speech compression, which reduces transmission time in mobile applications. They are used in denoising, edge detection, feature extraction, speech recognition, echo cancellation and others. They are very promising for real time audio and video compression applications. Wavelets also have numerous applications in digital communications. 1.1.2. Quantization

Quantization refers to the process of approximating the continuous set of values in the image data with a finite (preferably small) set of values. There are two types of quantization. They are scalar quantization and vector quantization. In

Fig. 1. Basic steps in wavelet-based image compression. 1550012-2 2nd Reading

April 17, 2015 13:51 WSPC/S0219-6913 181-IJWMIP 1550012

Empirical evaluation of EZW and other encoding techniques scalar quantization, each input symbol is treated separately in producing the output, while in vector quantization the input symbols are clubbed together in groups called vectors, and processed to give the output. 1.1.3. Encoding techniques

Coding is the process to remove the statistical redundancy in a given source. A coding algorithm producing an embedded code has the property that the bits in the bit stream are generated in order of importance, so that all the low rate codes are included at the beginning of the bit stream. Typically, the encoding process stops when the target bit rate is obtained.

Specialized encoding techniques are very good at compressing some types of information, like images or sound, but have poor results for other types of data.

Lossy image compression techniques provide much higher compression ratios than lossless image compression schemes. They are widely used since the quality of the reconstructed images is adequate for most applications. In this paper, lossy image compression encoding techniques are implemented. By these schemes, the decompressed image is not identical to the original image, but reasonably close to it. 1.2. Motivation and justification of the proposed approach