Abstract
Multimedia data, increasing in the modern era because multimedia are the major source of information. Multimedia data required storage capacity and transmission bandwidth. These factors we need for multimedia compression technique. Uncompressed data required more storage and transmission bandwidth on the other hand, we have limited storage capacity and transmission bandwidth. But improving in compression techniques, solve this problem. The primary focus of this paper to detail study of compression techniques.
Keywords
Compression, Decompression, Discrete Cosine Transform (DCT), Discrete Wevelet Tranfrom (DWT), Finite Ridgilet Transform (FRT), Image, Lossless,Lossy compression.
I. INTRODUCTION
The major reason for applying the compression techniques on data. We can reduce, storage requirement, processing time, and transmission duration. Compression allows a more efficient means to save memory capacity and transmission bandwidth and also the transmission duration. During the past years as compared the contributed of compression the fact that nowadays compression technology plays an important role in our life because it change the way of working on multimedia as image, video, text, audio, speech etc [1].
The basic purpose of compression is that we can beam data with minimal number of numbers according to human visual perception (HVP) [2]. A digital image has data contain are redundant and irrelevant information. For reducing redundancy we just focus on removing replication from the source of the image on the other hand, data signals which is omitted from source remove that signals is not received by the receiver such as human visual perception (HVP) is called irrelevancy reduction. Eliminating the redundant information from the image saving of storage space of an image.
The amount of digital data and resolution increasing day by day because standard reduces data, various amounts of data standard like MPEG-4, MPEG-7 and MPEG-21 after additional functionalities [3].
The three basic data redundancy namely spatial redundancy, which is defined spatial reducing image size by using a smaller number of bits and also correlation between neighbor pixel values, Spectral redundancy, Spectral which defining correlation between different color, Temporal redundancy, which is defining correlation between different frame in sequence of image [4].
The compression techniques, design for future challenges and advance application for multimedia and communication system. There are several compression techniques, but the main two categories are lossless and lossy compression techniques. Lossless compression there is no loss of information from image. Lossless gives 100% recovery data from the original data on the other hand the lossy compression technique gives a high compression ratio [5-9].
Figure 1: Research Trends of Multimedia Compression Techniques [10].
II. COMPRESSION ALGORETHM
Digital image compression is the major research areas due to continuously increasing various applications in different fields. This Section review the digital compression algorithm.
A LOSSLESS COMPRESSION TECHNIQUE
The need of lossless compression techniques is required in many applications such in medical data there were no loss of information. Lossless compression reduces storage space without degrading image and time needed for computational will also decrease. The Lossless image compression algorithm for different application like medical image, Lossless interframe coding for
MRI image, Ultersounsd image, Capsule Endoscopy (CE) image [11].
The different techniques that are used in lossless compression are Huffman coding, Shanan Fano Coding, Run Length Coding, Arithmatic Coding, Golomb Coding and Symbol-base Coding. Dictionary Algorithm such as LWZ, Area Coding, Bit plane Coding, Byte Pair Coding, Lossless Predictive Coding, Predictive Partial Matching act [12].
B LOSSY COMPRESSION TECHNIQUE
The lossy compression techniques give the high compression ratio then lossless, but the loss of data as compared to lossless compression. Lossy compression consists of three parts. The first is a transformation which gives a high compression ratio, the second part is quantization, which is reduced a symbol of bit represent the image, gives key issue which distinguishes between lossy or lossless compression methods. At the end used compression of entropy encoding [13].
Figure 2: Example of Lossy Compression Technique
III. DISCRETE COSINE TRANSFROM
Discrete cosine transform most suitable for medical image compression. A discrete cosine transform (DCT) is a sequence of finite data point in term of the sum by cosine function at different frequencies [14]. The two dimensional DCT is the essence of most popular lossy digital compression system today [15].
Figure 3: Example of Discrete Cosine Transform.
IV. 3D-DISCRETE COSINE
3D-DCT is used in image and video compression method for both JPEG and MPEG but these methods are not lossless. The three dimensional discrete cosine transform is used to produce a spectral frequency spectrum [16].
V. IMAGE COMPRESSION USING NEURAL NETWORK
A neural network image compression follows the following steps: store the color image for moderate size; for decomposing the discrete wavelet transform is used to the image for obtaining appromaxtion coefficient; for bit stream Huffman coding is used to compress image; by reverse process we obtain reconstruction.
VI. DISCRETE WAVELET TRANSFROM
Discrete Wavelet Transform is applied on de-noise image. Finite Ridgelet Transform (FDT) is used to obtain wavelet coefficient; compressed image of reduced sized is obtained; decompression is completed by applying an inverse Finite Ridgelet Transform (FRT) and Discrete Wavelet Transform (DWT) and the original image is obtained without loss of data [18].
This new technique for image compression give benefits for medical applications. Reducing computational complexity, mean square error, high compression ratio and better efficiency are obtained. The steps are: converted input image 256??256;converted RGB to gray; the third step is feature extraction is done; input image data is segmented and transformed to a set of features; for last stage decompressed image binary decoding is implemented [19].
VII. LITERATURE REVIEW
The literature study in the field of data compression are given below.
In [21], the author presented vector quantization based image compression technique. It can substantially improve the quality of vector quantization (VQ) compressed image. The vector quantization scheme is a lossy image compression for grayscale images. VC consists of three principal, codebook generation, image encoding and image decoding.
In [22], the author presented a lossless compression scheme for binary image. This method consists of two steps: first encode binary image using an encoding method than encode image data. Second is compressing the encoded data.
In [23], the author presented technique which is called five modulus method (shorty FFM) is consisted of dividing images into block 8*8 pixel each. We must known each pixel is a number b/w 0 to 255 for each of RGB array. After that the value could be divided by 5 to get a new value. Here we have a new formula to transform any number in the range 0-255 into a number that when divided by 5 the answer is always lying b/w 0-4.
In [24], the author presented a two dimensional differencing operation is first applied to image. The difference image is segmented and classified all black or all white or mixed blocks and group into a non overlapping region of all white and mixed blocks. non-overlapping region of the mixed block represented variable size segmentation and coding scheme.
In [25], the author presented a coding and decoding algorithm using a Matlab software is called Huffman coding scheme. They compressed image by reducing a bit per pixel as required a representative image. Image is reconstructed by using the decoding algorithm of Huffman technique.
In [26], the author presented Huffman coding techniques is used to compress files for transmission used statistical coding, Author said that Huffman coding is a the most frequently used symbols have shorter code word. Used for transmission a text and fax application that used Sarvel data structures. LWZ and Huffman, both used for compressed files, but LWZ takes more computational time therefore Huffman used to prefer. Huffman coding didn’t work well when the image is formed in binary than LWZ is used because this time gave a better compression ratio.
In [27], the authors presented a new technique which is used by the lossless grayscale image. This method work in two stages, first: analyze the set of model parameter in this way that the reduce the length of the encoded image. Second: second stage the coding stage is used to do the actual encoding.
In [28], the author presented introduced a new method for image compression, which is a combination of three techniques, namely as cryptography, multipath algorithm and steganography. These three are combined together along a Huffman algorithm to encrypt an image in an efficient way. By using Huffman the segmented image files are combined. We segmented the entire segmented image to compress into a single picture.
In [29], the author presented a bi-level image compression techniques using neural network. The multilayer perceptron neural network is applied that image pixel location. For encoded Huffman is used to encode and stored the compressed image. The production of this new technique comes out the pixel intensity 0 or 1.
In [30], the author presented a new concept for text compression.Transforming text character into a stream of words is not an easy process. This technique handles a bit-level as each character has its specific binary presentation.
In [31], the author presented a method based on Multi-table’s Arithmetic coding. Encode a 1s and 0s image. The method contain three approaches: median edge detector processing, base-switching transformation and statistical-model-based compressing.
VIII. OPEN AREA
The image compression method becomes increasingly important and new methods developed because of imaging data grow rapidly. Medical imaging is one is the best application area for the advance compression method. Another open field for future working is the developing of additional method based on set redundancy.Finally the use of lossless compression is increased because developing an enhanced compression model in the field of lossless compression technique. Another application is as follows.
A ENGINEERING
In many engineering areas, simulation with the supporter of computer is rapidly applied for test new design. The SRC method is used to extract the set of redundancy as compared to the general compression method.
B ASTRONOMICAL
Astronomers repetitively taking picture of the sky. These pictures are stored as digital image database. The enhanced compression method gives the other application area for astronomical imaging data based.
C INDUSTRIAL
Image technology is used in industry for quality control and sometimes used for monitoring through a robot vision application.
CONCLUSION
This paper presents a different compression technique these compression techniques are used in different application some of general used and some compression techniques are particular types of files. After studying all compression techniques we can conclude that lossless compression is more efficient as compared to lossy compression.