International OPEN ACCESS Journal Of Modern Engineering Research (IJMER) | IJMER | ISSN: 2249–6645 | www.ijmer.com | Vol. 4 | Iss. 5| May. 2014 | 37 | Performance analysis of Hybrid Transform, Hybrid Wavelet and Multi-Resolution Hybrid Wavelet for Image Data Compression H. B. Kekre 1 , Tanuja Sarode 2 , Prachi Natu 3 1, 3 (Sr. Professor, Assistant Professor , Department of Computer Engineering, MPSTME/ NMIMS University, India) 2 (Associate Professor, Department of Computer Engineering, TSEC, Mumbai University, India) I. INTRODUCTION In recent years, tremendous increase in internet usage has resulted in increased use of digital and multimedia data. Images are crucial part of this digital data as they represent the information effectively. Efficient storage manipulation and transmission of these digital images is equally important. Image compression serves this purpose. Aim of any compression technique is to eliminate redundant and irrelevant information while preserving the significant information [1]. This is possible as in case of images neighboring pixels are highly correlated and hence contain redundant information. Redundancy reduction removes the details in image which are not noticed by human visual system [2]. Data compression methods are usually classified into two categories: lossless compression and lossy compression. Image data compression is generally lossy compression. Lossy compression provides higher compression than lossless compression but decoded image approximately matches to original image [3]. Compression ratio is basic performance measurement criteria. It is defined as ratio of original data size and compressed data size. Higher compression ratio results in lower image quality and vice versa. Transform based coding is widely used for image compression in which image is transformed from spatial domain to frequency domain. Transforms have property to concentrate useful information into few low frequency coefficients. Initially Fourier transform was used for image compression. In Fourier transform local properties of signal are not detected easily. To overcome this drawback short time Fourier Transform (STFT) was introduced which is also called as window transform. It gives local properties at the cost of global properties [4]. The length of window limits the resolution in frequency. Discrete cosine Transform (DCT) [5] is popular transform used for image compression due to its good energy compaction property. Many compression systems use block based DCT where image is divided into blocks of uniform size and transform is applied on individual block. It does not take into account the discontinuities across the boundaries and results in degraded image. This is called as blocking effect [6]. Wavelets provide solution to these problems [7]. Wavelet transform gives time and frequency representation simultaneously [8]. Wavelet transform has higher energy compaction property. It is applied on entire image rather than on blocked image hence it eliminates blocking effect. Multi-resolution representation of an image is another most important characteristic of wavelet transform. Information contents of the image depend on the local variations of image intensity. Multi-resolution representation provides a hierarchical framework for interpreting the image Abstract: Compression of digital images play vital role in transmission of multimedia data. This paper presents application of hybrid wavelet transform in image compression. Multi-resolution property of Wavelet transform helps to analyze the information contents of image effectively. This property has been used in image compression application. Hybrid wavelet transform is generated using two different component transforms. Various sizes of these component transforms can be used. In this hybrid wavelet, global and local properties of component transforms are incorporated and hence are called bi-resolution analysis. Different levels of resolutions can also be included in generated hybrid transform. Hence It is called multi-resolution analysis and is applied on images. At each level of resolution number of components can be changed. It provides great flexibility to generate hybrid transform matrix. Image is compressed using hybrid wavelet, hybrid wavelet with multi-resolution and hybrid transform. Their performance is compared and it has been observed that hybrid wavelet transform gives lower error values than multi-resolution analysis and hybrid transform. Along with Root mean Square Error (RMSE), Mean Absolute Error (MAE) and Average Fractional Change in Pixel Value (AFCPV) is used to measure error. AFCPV gives better perception to image quality as it is a fractional change in pixel values. Lower the value of AFCPV better is the image quality. Keywords: AFCPV, Compression Ratio, Hybrid Transform, Hybrid Wavelet transform, Multi-Resolution Transform.