Search

In this paper, we propose a system-level error tolerance scheme for systems where a linear transform is combined with quantization. These are key components in multimedia compression systems, e.g., video and image codecs. Using the concept of acceptable degradation, our scheme classifies hardware faults into acceptable and unacceptable faults. We propose analysis techniques that allow us to estimate the faults' impact on compression performance, and in particular on the quality of decoded images/video. We consider as an example the discrete cosine transform (DCT), which is part of a large number of existing image and video compression systems. We propose methods to establish thresholds of acceptable degradation and corresponding testing algorithms for DCT-based systems. Our results for a JPEG encoder using a typical DCT architecture show that over 50% of single stuck-at interconnection faults in one of its 1D DCT modules lead to imperceptible quality degradation in the decoded images, over the complete range of compression rates at which JPEG can operate.

A wavelet image compression method using sub-block DCT is presented. The presented method is a hybrid type of DWT and DCT. The sub-block DCT is employed to the DWT-details. DWT-details have usually zero mean and small variance and hence the sub-block DCT method is very efficient. The size of sub-block for DCT is 8/spl times/8 pixel for level 1 DWT-details and 4/spl times/4 pixel for level 2 DWT-details. The simulation examples show that the presented method gives a better picture quality that of the usual JPEG method.

The major driving force behind the emerging of broadband integrated networks is traffic from video sources. Since this traffic will be a substantial portion of the overall traffic carried by the integrated networks, it is extremely important that adequate source models are found. The networks' dimensioning depends on the statistical properties of traffic sources. For video sources the Moving Picture Experts Group (MPEG) compression scheme was proposed several years ago and it has become the de facto standard for video compression since then. However, even with the huge reduction of bits that MPEG compression provides, it does not smooth the video traffic. Indeed the MPEG compression algorithm guarantees that the MPEG stream will be bursty. The aim of the work outlined was finding a suitable model for the MPEG coded video traffic. The paper concentrates on a variety of models and their capabilities to model adequately the statistical properties of MPEG video. In particular, the transform-expand-sample (TES), the autoregressive (AR), the autoregressive moving average (ARIMA) and the first order statistics histogram models were investigated in terms of both accuracy and complexity. A short discussion regarding the tradeoff between the accuracy and the complexity of these models is also provided.

As the applications of digital imagery expand in resolution and pixel fidelity there is a greater need for more efficient compression and extraction of images and sub-images. No longer is it sufficient to compress and decompress an image for a specific target device. The ability to handle many types of image data, extract images at different resolutions and quality, zoom and pan, and to extract regions-of-interest are the new measures of image compression systems. CREW (compression with reversible embedded wavelets) is a high quality image compression system that is progressive from high compression to lossless, pyramidal, and supports regions-of-interest and multiple image types. This paper describes the CREW system and format, shows how the correct data can be quickly extracted from a CREW file to support a variety of target devices, describes the mechanisms needed for panning, zooming, and fixed-size compression, and explains the superior performance on bi-level and graphic images.

Techniques are presented to progressively approximate and compress in a lossless manner two-colored (i.e. binary) 3D objects (as well as objects of arbitrary dimensionality). The objects are represented by a region octree implemented using a pointerless representation based on locational codes. Approximation is achieved through the use of a forest. This method labels the internal nodes of the octree as GB or GW, depending on the number of children being of type GB or GW. In addition, all BLACK nodes are labeled GB, while all WHITE nodes are labeled GW. A number of different image approximation methods are discussed that make use of a forest. The advantage of these methods is that they are progressive which means that as more of the object is transmitted, the better is the approximation. This makes these methods attractive for use on the worldwide web. Progressive transmission has the drawback that there is an overhead in requiring extra storage. Aprogressive forest-based approximation and transmission method is presented where the total amount of data that is transmitted is not larger than MIN(B,W), where B and W are the number of BLACK and WHITE blocks, respectively, in the region octree of the set of objects.

Traditional MPEG-2 video compression uses test model 5 (TM5) as the rate-control algorithm, which does not consider the perceptual quality of the compressed sequence. This paper proposes a new rate-control algorithm, which integrates TM5 with the Watson's digital-video-quality (DVQ) metric. As a proponent of the VQEG (Video Quality Expert Group), the Watson's DVQ is a subjective quality metric based on the human visual system (HVS) which estimates the perceptual quality of human viewers. Simulation results indicate that the proposed algorithm outperforms TM5 for low-motion video sequence, and performs about the same as TM5 for high-motion video sequences.

This paper describes a methodology for obtaining watermarking schemes using different compression algorithms in the mark embedding process. The compression methods enclose both spatial and transformed domains and both scalar and vector quantization. The result is a family of watermarking algorithms which are robust against different lossy image compression, filtering, cropping and row/column removal attacks, for a reasonable range of image qualities. In addition, the experiments show that the JPEG2000 standard is a promising tool for creating new watermarking schemes.

Block matching algorithms (BMA) have been widely used as standard motion estimation (ME) techniques in video coding practice. In this paper, we present an iterative but analytical motion estimation algorithm with attractive computational efficiency. Our work can be viewed as generalization of the celebrated work of Lucas and Kanade (1981) on iterative image registration into the wavelet space. We demonstrate that block-based motion estimation can be performed efficiently in a coarse-to-fine fashion by propagating the estimate across the scale in the wavelet space. Our solution often takes a fraction of the computations required by the conventional BMA methods and reaches comparable MSE performance of prediction residues (fractional-pel motion accuracy is achieved due to the analytical nature of the proposed scheme). In spite of the iteration in the proposed scheme, it converges fast especially with a good initial estimate passed on from the low resolution. Our new ME scheme is particularly useful to the speed-up of implementing a new class of wavelet-based scalable video coders.

This paper addresses the vectorization of the lifting-based wavelet transform on general-purpose microprocessors in the context of JPEG2000. Since SIMD exploitation strongly depends on an efficient memory hierarchy usage, this research is based on previous work about cache-conscious DWT implementations. The experimental platform on which we have chosen to study the benefits of the SIMD extensions is an Intel Pentium-4 (P-4) based PC. However, unlike other authors, the vectorization has been performed avoiding assembler language programming in order to improve both code portability and development cost.

In this paper, we propose an optimized real-time MPEG-2 video decoder. The decoder has been implemented in one FPGA device as a HW/SW partitioned system. We made timing power-consumption analysis and optimization of the MPEG-2 decoder. On the basis of the achieved results, we decided on hardware implementation of the IDCT and VLD algorithms. The remaining parts were realized in software with a 32-bit RISC processor. The MPEG-2 decoder (RISC processor, IDCT core, VLD core) has been described in high-level Verilog/VHDL hardware description language and implemented in Virtex 1600E FPGA. Finally, the decoder has been tested on an industrial prototyping board.

This paper discusses the application of a new two-dimensional algebraic integer encoding scheme for the design of a processor core for JPEG and MPEG applications. The paper concentrates on the efficient implementation of the 2D algebraic integer DCT. The processor takes advantage of the low complexity, multiplierless, high-precision nature of the algebraic integer encoding scheme. Details are provided for the DCT processor including a description of 0.18 /spl mu/m CMOS test chip recently submitted for fabrication.

Naming image compression's state of the art surely leads to JPEG2000. This new format takes advantage of the discrete wavelet transform to overtake all the compression methods that have been used up to now. This article proposes two changes in the processing chain to deal with sparse histogram images: one nested in Tier-1 coding, the other between coefficient coding and MQ-coder.

In this paper, we propose an optimized real-time MPPEG-2 video decoder. The decoder has been implemented in one FPGA device as a HW/SW partitioned system. We carried out time/power-consumption analysis and optimization of the MPEG-2 decoder. On the basis of the achieved results, we decided for HW implementation of the IDCT and VTD algorithms. Remaining parts were realized in SIV with 32-bit RISC processor. MPEG-2 decoder (RISC processor, IDCT core, VLD core) has been described in Verilog/VHDL and implemented in Virtex 1600E FPGA. Finally, the decoder has been tested on the Flextronics prototyping board.

In this paper, a "Circuit-sharing scheme" is proposed to design a reduced chipset for DAB (digital audio broadcasting) receiver, involved the fast Fourier transform (FFT) circuit in the orthogonal frequency division multiplexing (OFDM) and IMDCT (inverse modified discrete cosine transform) synthesis filter in the audio decoder Comparing to the existent method, it can achieve a highly integrated, low power and small size DAB receiver. We implement the IMDCT synthesis filter of MPEG audio decoder and the 2048-point complex FFT by a single processing element with pipeline structure. Simulations demonstrate that the proposed circuit sharing can reduce the chipset area and meet the specification of DAB. The gate count is about 6692 excluding the 2048/spl times/16 RAM and 512/spl times/32 ROM.

Summary form only given. The approximate pattern matching on the text transformed by the Burrows-Wheeler transform (BWT) was considered. This is an important first step towards developing a compressed pattern matching algorithm for the BWT based compression system. Algorithms are proposed to solve the K-mismatch problem. Tests were performed on different pattern lengths using 133 selected files from the Canterbury, Calgary, and TREC corpus. The results on the K-mismatch pattern matching show that the running time and storage are superior to the fast suffix tree approach. Thus, once the index arrays are created, for repeated pattern search operations and for long patterns, the proposed algorithms perform significantly better than the agrep and ngrep. Using DFA verification, the search time is almost constant. The amortized cost is lower for multiple patterns search operations.

Summary form only given. The problem of efficient image/video transmission over wireless channels is considered. Such a problem involves good compression rates and effectiveness in presence of channel failures. The proposed work includes a balanced multiple description-coding (MDC) schemes for still image based on the discrete wavelet transform (DWT) developed. This MDC includes an efficient bit allocation procedure that dispatches the source image redundancy between different channels. Extended version for video was presented, and uses the 3D scan-based DWT that involves scan-based MDC with rate or quality control. To solve this problem, redundancy between the descriptors in function of channel model and state (BER) is proposed.

Summary form only given. StarZip, a multi-copora text compression system, was introduced together with its transform engine StarNT. One of the key features of the StarZip compression system is to develop domain specific dictionaries and provide tools to develop such dictionaries. StarNT was utilized because it achieves a superior compression ratio than almost all the other recent efforts based on BWT and PPM. StarNT is a dictionary-based fast lossless text transform. The main idea is to record each English word with a representation of no more than three symbols. This transform maintains most of the original context information at the word level and provides an "artificial" strong context. It ultimately reduces the size of the transformed text that, in turn, is provided to a backend compressor. This data structure provides a very fast transform encoding with a low storage overhead. StarNT also treats the transformed codewords as an offset of words in the transform dictionary. The time complexity for searching a word in the dictionary is achieved in the transform decoder. Experimental results have shown that the average compression time has improved by orders magnitude compared to previous dictionary-based transform LIPT. The complexity and compression performance of bzip2, in conjunction with this transform, is better than both gzip and PPMD. Results from five copora have shown that StarZip achieved an average improvement in compression performance (in terms of BPC) of 13% over bzip2-9, 19% over gzip-9, and 10% over PPMD.

Summary form only given. Modern image compression standards such as JPEG 2000 are only defined to operate on integer data and, thus, they pose a problem if data need compressing. The typical way to handle floating point data is to do an initial quantization. This approach allows standard techniques to be used on the quantized data and may be acceptable to researchers due to the irreversible data loss. Another compression approach is through the use of strictly lossless methods such as LZW based techniques. The disadvantages of this approach include low compression ratios, no capability for lossy compression, and low flexibility of the compressed data stream. Extensions to the JPEG2000 standard were proposed that allow efficient floating point data coding. These extensions have the advantage of maintaining desirable properties of JPEG2000, including flexible embedded bit streams, lossy and lossless compression image scalability. Bit plane coding is needed for an overall floating point algorithm and can be applied when acceptable wavelet and color transforms are derived. The derivation of the floating point bit plane coding algorithms considers only the narrow problem of JPEG2000. The proposed approach for bit plane coding represents floating point values as "big integers" and codes these integers directly as in JPEG2000. A separate eligibility pass every several bit planes was introduced to lower the number of coefficients coded in each coding pass resulting in improved efficiency.

Summary form only given. Transport of reversible and unreversible embedded wavelets (TRUEW) is a system for the portion delivery of compressed JPEG 2000 files over a network. A discussion on the development of the TRUEW architecture, the benefits of the TRUEW architecture, an efficiency analysis of tile-part and precint access to JPEG 2000 compressed images, and some comments on potential applications of interactive image access are presented. Computations show that for an example access pattern, TRUEW provides comparable performance in terms of disk accesses and data transferred with more complex precinct based systems.

Orthogonal transforms are compared with the casual transform in lossless transform coders. For single-stage lossless coding, it was previously shown that the integer-to-integer implementation of the best orthogonal decorrelating transform, the KLT, leads to lower compression performance than its casual counterpart. The analysis in the framework of a multi-stage lossless coding scheme, which yields a lossy coded signal, and an error signal was pursued. This scheme allows one to choose the respective bit rates of both complementary signals, depending for example on the bandwidth of the transmission link. It was shown that the casual approach presents several advantages w.r.t. its orthogonal counterparts. For orthogonal transforms, the price paid for the multiresolution approach is a bit rate penalty of 0.25 bit per sample. This excess bit rate is due to a "gaussianization effect" of the transforms. Firstly, it was shown under the assumptions of smooth p.d.f.s for the sources, and of high resolution for the lossy coded signal, that the casual approach allows one to code the data without causing any excess bit rate as compared with a single-stage coder. Secondly, the approach based on the casual transform allows one to easily switch between a single- or a multi-stage compressor. Thirdly, in the framework of interchannel redundancy removal, this approach allows one to easily fixed the distortion and rate for both the low resolution and the error signal of each channel, by using different stepsizes in the quantization stage. Any of the channels may, as a particular case, be chosen to be directly losslessly coded. Finally, a side advantage of the casual approach is that entropy coding of the error signal is made very simple since for odd quantization stepsizes, the discrete error sources are uniformly distributed, so that the optimal codewords have the same length, and fixed rate coding is optimal.

Summary form only given. The Burrows Wheeler transform is a relatively new approach to text compression and has already proven to produce excellent results. However, there has been much research directed towards improving the efficiency of the move to front algorithm with varying degrees of complexity. A relatively simple technique was examined using a dual modeling system and achieves very promising results. Information loss during the Burrows Wheeler transform is also examined using deterministic information theory measuring techniques. By utilizing the information measurement tools developed by Titchener a clearer understanding of informational changes over BWT has been established. Most importantly, MTF loses information making it amenable for compression, however it also contains little contextual structure and closely resembles the data stream after applying the BWT algorithm.

Summary form only given. A new algorithm, referred to as the wavelet packet-based embedded blocking code (WPEB), was developed for synthetic aperture radar (SAR) data compression. This algorithm combines the different properties of wavelet packet decomposition, block coding, and speckle reduction. Better results were obtained by using the wavelet packet transform that decomposes the higher frequencies. The higher coding efficiency can be obtained by dynamically allocating bits according to block variances. The examples using RADARSAT data have shown that compression performance is better than conventional wavelet methods. Visual image interpretation was found acceptable at 1 bpp.

Many image coders map the input image to a transform domain, and encode coefficients in that domain. Usually, encoders use previously transmitted samples to help estimate the probabilities for the next sample to be encoded. This backward probability estimation provides a better PDF estimation, without need to send any side information. A new method of encoding was proposed that goes one step further: besides past samples, it also uses information about the current sample in computing the PDF for the current sample. Yet, no side information was transmitted. The initial PDF estimate is based on a Tarp filter, but probabilities are then progressively refined for non-zero samples. Results are superior to JPEG2000, and to bit-plane Tarp.

Summary form only given. Hyperspectral images were generated through the collection of hundreds of narrow and contiguously spaced spectral bands of data producing a highly correlated long sequence of images. An investigation and comparison was made on the performance of several three-dimensional embedded wavelet algorithms for compression of hyperspectral images. These algorithms include 3D-SPIHT, AT-3DSPIHT, 3D-SPECK (three-dimensional set partitioned embedded block), and JPEG2000. 3D-SPIHT is the three-dimensional extension of the state-of-the-art SPIHT (set partitioning in hierarchical trees) image compression algorithm. AT-SPIHT uses the same coding algorithms as 3D-SPIHT, but utilizes a more efficient, asymmetric tree structure. 3D-SPECK is a version of two-dimensional SPECK modified to encode 3D subband blocks. For an image sequence, 3D-DWT was applied to obtain a wavelet coefficient prism. This prism is partitioned into small code blocks with different sizes, and each subband is treated as a code block. A block splitting algorithm is used on the individual code blocks to test their significance. Compression performance was conducted on the mentioned algorithms using several AVIRIS (Airborne Visible Infrared Imaging Spectrometer) image sequences. The rate distortion performances were compared by means of the peak signal-to noise ratio (PSNR). The visual effect is considered a primary concern of hyperspectral imagery so an analysis of every algorithm's visual effect is in order. Through experimentation, AT-3DSPIHT provided high quality reconstructed images even at very low bit rate such as 0.4 bpp. A comparison was also conducted on JPEG200 and other algorithms concerning JPEG2000's multi-component images. The results have shown that AT-3DSPIHT, 3D-SPIHT and 3D-SPECK outperform JPEG2000 by the approximate range of 1 to 2.5 dB PSNR, depending on the image sequence.

The error resilience properties of JPEG2000 are investigated. Through a careful analysis of the structure of the JPEG2000 codestreams, the dependencies among coding passes of a codeblock are identified. In the analysis, the effects of mode variations provided by the standard for low-memory, low-complexity implementations are considered. The proposed methods are derived using the existing dependency structure of coding passes and do not require a substantial increase in computational capabilities of the decoder. Experimental results indicate that these methods can improve error resilience performance substantially.

This paper presents a systolic realization of the discrete wavelet transform used in JPEG 2000 standard on Systola 1024 computer. A general idea of image compression, JPEG and JPEG 2000 standards, is explained.

Summary form only given. Data storage issues pose a potential problem in the performance and analysis of turbulence simulations. One tool that has successfully addressed such performance problems in other domains is data compression. Unfortunately, data from turbulence simulations does not usefully compress with lossless methods. In this work, the compression of data from a simulation of barotropic turbulence using wavelet-based lossy coding is examined. The resolution of the data used is 512 by 512. The output of this simulation is the relative vorticity as single precision floating point values. A number of quantities are computed from the relative vorticity. The numerical error in these quantities is evaluated along with visualizations of the vorticity and error correlation with the uncompressed data. It is found that depending on the quantities of interest and the evaluation criteria, compression ratios of 4:1 to 256:1 are achievable. Under a conservative definition of acceptable error, it is possible to recover quantities of interest from data compressed 4:1 (8 bpp), the data rate that in existing practice is used for visualization.

The video compression algorithms based on the 3D wavelet transform obtain excellent compression rates at the expense of huge memory requirements, which drastically affect the execution time of such applications. The goal of this work is to mitigate the memory problem by exploiting the memory hierarchy of the processor through blocking. In particular, we present two blocking approaches: cube and rectangular that differ in the way that the original working set is divided. We also propose the reuse of previous computations in order to decrease the number of memory accesses and floating point operations. Results show that the rectangular overlapped approach with computation reuse obtains the best results in terms of execution time, a speedup of 2.42 over the non-blocking non-overlapped wavelet transform, maintaining the compression ratio and the video quality (PSNR) of the original encoder based on the 3D wavelet transform.

By reducing the accuracy of the logarithmic number system (LNS) it is possible to achieve lower power consumption for multimedia applications, such as MPEG, without significantly lowering the visual quality of the output. An LNS wordsize of 8 to 10 bits produces a comparable MPEG output as a fixed-point wordsize of 14 to 16 bits. The switching activity of an LNS ALU that computes the inverse discrete cosine transform (IDCT) is one quarter that of fixed point, implying lower power consumption. By skipping inputs that are zero (which MPEG can do naturally with its run-length coding and zigzag ordering) the switching activity of LNS MPEG becomes one-tenth that of fixed point, in contrast to the minimal impact zero skipping has on fixed-point power consumption.

We propose an algorithm for image recovery where completely lost blocks in an image/video-frame are recovered using spatial information surrounding these blocks. Our primary application is on lost regions of pixels containing textures, edges and other image features that are not readily handled by prevalent recovery and error concealment algorithms. The proposed algorithm is based on the iterative application of a generic denoising algorithm and it does not necessitate any complex preconditioning, segmentation, or edge detection steps. Utilizing locally sparse linear transforms and overcomplete denoising, we obtain good PSNR performance in the recovery of such regions. In addition to results on image recovery, the paper provides further insights into the usefulness of popular transforms like wavelets, wavelet packets, discrete cosine transform (DCT) and complex wavelets in providing sparse image representations.

In this paper, we analyze the performance of DFT codes with bursty erasures in the framework of syndrome decoding. Bursty erasures give rise to a syndrome decoding matrix which has very large elements depending on the code parameters and the burst length. The largeness of the elements of the syndrome decoding matrix is studied by establishing various relationships between the syndrome decoding matrix, and the parity and the generator polynomial coefficients. With a suitable model for the quantization error, the reconstruction error performance of DFT codes in the context of the syndrome decoding is analyzed and then applied to the case of bursty erasures. Simulation results with a Gauss-Markov source verify the theoretical results obtained with the assumed quantization error model.

Summary form only given. We present a detailed description of a lossless compression algorithm intended for use on files with non-uniform character distributions. This algorithm takes advantage of the relatively small distances between character occurrences once we remove the less frequent characters. This allows it to create a compressed version of the file that, when decompressed, is an exact copy of the file that was compressed. We begin by performing a Burrows-Wheeler (1994) Transform (BWT) on the file. The algorithm scans this BWT file to create a character frequency model for the compression phase. To deal with the issue of bit encoding, we write every number as a byte or sequence of bytes to the compressed file and run an arithmetic encoder after the file has been compiled.

A progressive method for transmission of images over a bursty noise channel is presented. It is based on discrete wavelet transform (DWT) coding and channel-optimized scalar quantization. The main advantage of the proposed system is that it exploits the channel memory and hence has superior performance over a similar scheme designed for the equivalent memoryless channel through the use of channel interleaving. In fact, the performance of the proposed system improves as the noise becomes more correlated, at a fixed bit error rate. Comparisons are made with other alternatives which employ independent source and channel coding over the fully interleaved channel at various bit rates and bit error rates. It is shown that the proposed method outperforms these substantially more complex systems for the whole range of considered bit rates and for a wide range of channel conditions.

On-line adaptation to nonstationary distributions is essential to good performance in image coding. Fixed-size contexts (with adaptive tables) are also widely used, in conjunction with arithmetic encoders, in state-of-the-art codecs. In contrast, we propose a simple two-dimensional filter that directly outputs the probability distribution function (PDF) estimate necessary to drive an adaptive arithmetic encoder. The filter is isotropic, in the sense that the impact of a previously encoded bit depends only on its value and distance to the bit to be coded. Surprisingly, this simple filter yields results comparable to or better than JPEG2000. It also brings an interesting distinction between on-line and off-line learning, and their relative importance in compression.

Summary form only given. An approach based on applying Delaunay triangulation to compression of mean values of image blocks that have non-identical shape and size is proposed. It can be useful for image compression methods that require the use of image partition schemes with non-equal block size like fractal and DCT-based image coding. Several methods of block mean value coding are considered. In particular, the drawbacks of using quantization with further redundancy elimination by entropy coders are discussed. Another considered method is the forming of the block mean value image and its further compression by lossy coders. Finally, the motivations in favor of Delaunay triangulation application to block mean value image coding are presented.

We are concerned with multiple description coding of predictively encoded sequences at low rates/redundancies. The key problem to be solved is that of avoiding predictive mismatch. We propose an algorithm based on multiple description correlating transforms (MDCT), which avoids the problem of predictive mismatch while not increasing the number of transmitted coefficients. Under a general first-order, Gauss-Markov source model, carefully formulated to be applicable to the practically important case of video coding, we obtain closed-form expressions for the operational rate-distortion characteristics of the proposed algorithm and use these to optimize the algorithm encoding parameters. In the case of the residual-of-residuals technique, which is also based on MDCT, we show that incorporation of any redundancy between the transmitted descriptions is inherently suboptimal. Results show that the proposed algorithm significantly outperforms the residual-of-residuals technique at low rates.

This paper proposes a compressed domain video watermarking scheme for copyright protection. Our scheme is designed based on the concept of communications with side information. For making the real-time detection a reality, the watermark is directly embedded and detected in the VLC domain. The typical problems of video watermarking such as preservation of bit rate, video attacks, real-time detection is examined. The performance of the new watermarking scheme is examined by checking its robustness capability against attacks together with false positive analysis.

In this paper, the impact of scene change on the conventional error concealment method is addressed and a novel error concealment method is proposed to improve the insufficiency of the conventional temporal error concealment algorithm due to the occurrence of scene changes. Combined with the low complexity scene change detection algorithm using macroblock type information, the corrupt blocks resulting from bit errors are concealed either temporally or spatially depending on whether or not an abrupt scene change is found. In the case of a gradual scene change, a novel error concealment method of interpolation and extrapolation is proposed to utilize the linear property of a gradual scene change sequence, and effectively reduce the concealment error compared with the conventional algorithm. A great improvement of about 3 to 5 dB PSNR is obtained with only small memory and computation overhead.

We have implemented a client/server system for fast trial-and-error analysis: the StreamCorder. The server streams wavelet-encoded views to the clients, where they are cached, decoded and processed. Low-quality decoding is beneficial for slow network connections. Low-resolution decoding greatly accelerates decoding and analysis. Depending on the system resources, cached data and analysis requirements, the user may alter the minimum analysis quality at any time.

Spatially varying sensing (foveation) was first used as a means for image compression in our past research (1993). In this report we extend previous work by us to address the advantages of foveation in improving the performance of interactive 3D visualization over bandwidth limited channels, such as the Internet. We develop an algorithm that is implemented in JAVA3D. The method combines foveating JPEG texture files with level-of-detail representation in JAVA3D. Applications of the work include model based coding following the Synthetic-Natural Hybrid (SNHC) paradigm in MPEG-4 to Electronic Commerce with 3D visualization of merchandise. Experimental results are presented to validate our approach.

The theory of projection onto convex sets (POCS) is applied to reduce blocking artifacts in compressed images coded by the block discrete cosine transform (BDCT). An image before compression is simulated by a triangular mesh, which is the basis of a proposed smoothness constraint set. The mesh is constructed by dividing each block into a set of triangles. The proposed method outperforms four existing methods subjectively and objectively.

Summary form only given. This paper describes an adaptive error recovery technique for a DWT-based coding and transmission system, which recovers singularity information as well as continuity. The approach provided in this paper not only interpolates from the boundary, but also maintains the sharpness in background instead of smoothing the discontinuity out. The recovery function is modeled by the numerical solution of a special category of elliptic equation. The difference between adjacent data can be adjusted by the given weights. The choice of the weights allows high flexibility in maintaining contrast, since we can give small weights where we have evidence for edges.

Video streaming on the Internet often suffers from high frame loss rates due to fragmentation of large frames and inter-frame dependencies needed for high compression. We propose adding lightweight, priority-based queue management to Internet routers to significantly improve performance of video streaming. We extend a Rate-Based RED approach to support three priority classes and apply it to MPEG. The performance of the mechanism on video streams is measured, analyzed and compared with Drop-Tail and RED in terms of transport layer throughput, system fairness, application layer throughput and goodput, and video stream quality. Extensive simulation shows that our approach improves MPEG video stream quality and network system fairness over traditional RED under a variety of workloads.

In this paper we consider the problem of quantifying target distortion in lossy compression using the JPEG2000 Region Of Interest (ROI) option. This approach brings about new problems that need to be addressed, such as automated selection of target regions and ROI preservation parameters. By combining variations of traditional metrics and newly developed segmentation-based image quality metrics, we illustrate several ways to quantify target distortion using ROI preserving compression for various methods including JPEG2000.

Wavelet-based image coders optimally represent smooth regions and isolated point singularities. However, wavelet coders are less adept at representing perceptually important edge singularities, and coding performance suffers significantly as a result. We propose a novel two-stage image coder framework based on modeling images as edge cartoons + textures. In stage 1, we infer and efficiently code the edge information from the image using a multiscale wedgelet decomposition. In stage 2, we code the residual, "edgeless" texture image using a standard wavelet coder. Our preliminary coder improves significantly over standard wavelet coding techniques in terms of visual quality.

The main idea in SPEAR is to have a number of European experts work together on designing the new standard for coding and compression of still pictures, JPEG 2000. JPEG stands for Joined Photographic Experts Group, the ISO working group referenced as ISO/IEC JTCI SC29 WGI. JPEG 2000 is referenced as ISO 15444. It is due to become a standard by, the end of the year 2000, with implementations in a large,lumber of domains, thanks to its concept of "Open Standard" which gives a new dimension in terms of flexibility and functionality.

This paper presents a new methodology which performs both lossless compression and encryption of binary and gray scale images. The compression and encryption schemes are based on SCAN patterns generated by the SCAN methodology. The SCAN is a formal language based two dimensional spatial accessing methodology which can efficiently specify and generate a wide range of scanning paths or space filling curves. This paper presents compression specific SCAN language, compression and decompression algorithms, encryption and decryption algorithms, and test results of the methodology.

With the increasing use of multimedia technologies, image compression requires higher performance as well as new functionality. To address this need in the specific area of still image encoding, a new standard is currently being designed: JPEG2000. We see what the exact goals are of this future standard, which applications it addresses and the current standardisation process. Then we show through the descriptions of two available demonstrations what kind of results are already reachable with the current status of the standard. The first demonstration describes a Java implementation of the future standard, details the advantages of such an implementation and compares the performance of JPEG2000 with that of JPEG. The second demonstration describes how JPEG2000 can be used in domains were the transmission bandwidth is very restricted, taking advantages of new functions such as the definition of regions of interest and progressive transmission.

The performance of compression algorithms is traditionally evaluated by statistical metrics such as mean square error (MSE) and peak signal-to-noise ratio (PSNR). However it is well known that these metrics do not always correlate with visual perception. We have implemented a well known perceptual metric and evaluated the performance of a number of recently developed high fidelity compression algorithms using perceptual as well as statistical metrics. Our preliminary results do not indicate a consistent correlation between the perceptual metric used and subjective ranking.

Presents a fast rate control scheme that regulates the coded bit stream generated by an MPEG video encoder such that the resulting variable-rate compressed bit stream can be transmitted over a fixed-rate channel. A new bit-rate regulation scheme is proposed using a technique called pruning, in which only the required number of low-frequency coefficients of the discrete cosine transform (DCT) are computed, without the need to compute all the coefficients. The use of the pruning technique reduces the overall execution time of the encoder implementation.