2. REVIEW OF LITERATURE

Transcription

1 2. REVIEW OF LITERATURE Digital image processing is the use of the algorithms and procedures for operations such as image enhancement, image compression, image analysis, mapping. Transmission of information within document images was needed. This often required an analogue to digital conversion. This in turn has brought about the desire to be able to process the images in digital form and more importantly, process the content of these images GENERAL COMPRESSION ALGORITHMS Digital image compression continues to be an active area of research as a result of their increasing application in various fields. This section reviews general image compression techniques Lossy Techniques The nature of the human eye s perception allows significant reduction of information in the image for JPEG2000. The transform coefficients are rounded and quantized causing partial loss of information. These algorithms are optimized for compression of photographic images, which are mostly used in computer industry. There are also transform-based algorithms optimized for different tasks such as Enhanced Compression Wavelets for the compression of aerial and satellite photos. DCT and especially wavelet -based algorithms present excellent compression efficiency in terms of compression vs. degradation tradeoff for the class of images to which they were optimized. In some applications, it is not necessary to transfer the whole image data in one continuous transmission. It is often more important to have a schematic 31

2 thumbnail of the image faster than the whole image. This requirement is typical for browsing and retrieval applications in restricted bandwidth transmitting channels, when one first. Each step then updates the data finally giving the exact lossless Lossless Techniques Lossless data compression is a technique that allows exact reconstruction of the original image from its compressed form. Lossless compression is used when it is important that the original and the decompressed image states that it is necessary for many high performance applications such as geophysics, telemetry, nondestructive evaluation, and medical imaging, which other types of decorrelation techniques. The second stage, which includes Huffman coding, arithmetic coding, and LZW, removes coding redundancy. The performances of entropy coding techniques are very close to its theoretical bound, and thus more research activities concentrate on decorrelation stage. Image is 5 MB, the compressed image would be 1 MB. Authors note that, this may seem like a substantial reduction until one considers that irreversible compression techniques can produce compression ratios of up to 100:1, thus reducing the size of rather than the pixel values that make up the image. One such coding technique is run length encoding (RLE). In RLE, any item that is repeated in the image is another method of lossless compression that can operate on text data as well as on images. This method compresses data by first constructing a table of the relative frequency of the elements of the data. Then, the that at the beginning of the file, be substituted for those patterns, resulting in greater compression. It does this in one pass, resulting in greater speed. LZW can compress images up to 10:1. The typical digital formats TIFF (Tag Image Redundancy removal, or effective and standardized image compression techniques. Among the emerging standards are JPEG, for compression of still images (the set are the most meaningful; the latter, the 32

3 least. The least meaningful frequencies can be stripped away based on allowable resolution loss. DCT-based image compression relies on two techniques to associated with each pixel in the image. Entropy coding is a technique for representing the quantized coefficients as compactly as possible. color images. The algorithm started with a conversion from RGB color space to YCbCr color space, followed by DCT transform. Using Bisection method, an iterative phase including thresholding, quantization was performed to compress. A reverse process reconstructed the original image. The efficiency of the system was demonstrated through resutls and was compared to a block truncation -based coder. During the same, by subtracting, at the encoder, the low-quality image from the original image, obtaining the noiselet values and subjecting them to quantization and entropy coding. Fractal coding is a potential image compression scheme, which has the advantages of, and only the domain blocks with the same class to the range block are calculated during best match exploiting process. Experimental results showed that compared with standard fractal coding scheme, the encoding time is significantly reduced and the PSNR of the reconstructed image is also improved. bit plane encoding, and reversible integer pre- and post-filters. Simulation results show that the method is competitive against JPEG-LS and JPEG2000 in lossless compression and outperformed JPEG2000 in lossy compression. This section has reviewed only the recent works published on DCT. Many -based compression is another algorithm that is most frequently used by public and hence taken for analysis in the present research work Wavelet Image Coding Algorithms using wavelet technology which includes wavelet with vector quantization, embedded zerotree wavelet algorithm, Set 33

4 partitioning in hierarchical. JPEG uses both Huffman and run length encoding schemas for compressing images to reduce the volume of data. The JPEG cause objectionable 'ringing' around sharp edges, especially text. To overcome these problems, wavelet compression techniques (lossy ithms (eg, Discrete Wavelet Transform) are based on the use of high- and lowass filters coupled the bits in the bitstream in order of importance, yielding a fully embedded code. With an embedded bit stream the reception of code bits can be stopped at any point and the image can still be reconstructed. The algorithm is based on four key concepts, 1) a DWT performed in such a way that no child is scanned before the parent. During the scan, each component that is not marked insignificant by another zerotree is coded with one of four symbols (1) root (2) isolated zero (3) positive significant in the decomposition. These are the List of Insignificant Pixels (LIP), List of Significant Pixels (LSP) and List of Insignificant Sets (LIS). The LIP contains coordinates of components that are insignificant at the current threshold. The LSP contains the coordinates of components that are significant to the same threshold. ell with compound images. Because of this reason, researches for compound compression were dealt separately. Research focusing on Computer Generated Image (CGI) has captured the llows for compression of halftones. The DjVu In this approach, a full segmentation of the image is done to separate the different s, which increases the memory access times considerably, making real-time video compression very expensive or impossible. (b) Multi-layer image decomposition 34

5 The main appeal of this approach is that it is the shortest path to supporting compound image compression with existing standards. The idea is to decompose the images in different layers, some containing the color information, and some This technique is by far the simplest. The compound image is divided in blocks of a certain size (like 8x8), and a classification method is applied to the block to decide which compression method is to be used to code its pixels. The classification can be always avoided with sufficiently conservative classification rules. 2.3 LAYER AND BLOCK-BASED COMPOUND IMAGE COMPRESSION Research attempts in the field of compound image compression are generally classified as either layer-based or block-based Layer-Based Approaches Most of layer-based coding algorithms use the standard three-layer Mixed Raster Content (MRC) format. The main function of the algorithm is to divide a compound image into multiple regions, that is a text region and a nontext region. He used lossless compression for the text layer and lossy compression for non-text layer. Similarly, Said and Drukarev (1999) divided the compound images into background, color foreground and text. They used lossy JPEG for background, color tag representation for foreground and a token-based lossless approach for text region. Compressing the efficiently to match the compound layer characteristics. Their work also used block -based compression techniques (explained below) for compressing color compound images. A method for checking image compression using layered coding method is presented by efficiency of IW44 and JB2 heavily draws on their use of an 35

6 efficient binary adaptive arithmetic coder called the ZP-coder. The ZP-coder compares favorably with existing approximate arithmetic coders both in terms of compression ratio and speed. Mixed Raster Content (MRC) Model for Compound Image Compression sed approaches own very flexible coding structure. However, the coding performances are not good enough for both natural and compound images. Wong et al. (1982) proposed a technique called the Run Length Smoothing Algorithm (RLSA) to partition a binary compound image into blocks. Each block was then classified as he algorithm failed with hybrid block, that is, image containing mixed text and pictures. Several other approaches were also developed is the most important step in any research. Different researchers use different parameters for analysis. In general, almost all the The compression ratio can be measured as the ratio of the number of bits required to represent the image before compression to the number of bits required to represent the same image after compression. The compression ratio is equal to the compression technique employed is more effective. The tradeoff between compression ratio and picture quality is an important point to consider when compressing images. Speed of Compression.Deviation, Mean Square Difference, Equivalent Number of Looks, Deflection Ratio and Figure of Merit. 36