Abstract This paper emphasizes the joint time‐frequency interpretation of eigensubspace representation for network statistics as features for identification and tracking traffic flows based on the link level activity. Eigencoefficients (frequency domain feature) and eigenvector components (time domain feature) are jointly utilized to quantify their combined significance on the representation of each link data (each component of the link traffic vector) in the eigensubspace . The joint time‐frequency method is employed to analyze the traffic data obtained from the Internet2 network. It is shown that the analysis with link‐level resolution brings advantages for network traffic engineering applications. A machine learning method is investigated to identify network paths using eigenanalysis of link statistics as the feature set. The merit of the method is validated by experimental studies of the network scenarios considered in the paper. Eigenvectors and eigenflows in the subspace are jointly used as factors (features) for linear regression to forecast the network link traffic. It is demonstrated that the eigensubspace based auto‐regressive order two, AR (2), predictor is superior to the time‐domain based predictor to forecast the link level traffic of a network.

The high peak‐to‐average power ratio (PAPR) of the transmitted signal is a major drawback of multicarrier transmission such as orthogonal frequency‐division multiplexing OFDM. A plethora of PAPR reduction techniques has been reported in the literature. Some of the techniques modify the phase and/or amplitude of symbols in the original symbol alphabet (SA) such as selected mapping and partial transmit sequences techniques. However, such methods have shortcomings of a heavy computational burden caused by required multiple inverse fast‐Fourier transform (IFFT) operations and bit error rate performance degradation due to side information (SI). In this study, a low‐complexity PAPR reduction framework is proposed to jointly modify phase and amplitude values of the original symbols in the alphabet. This framework utilises only one IFFT/FFT operator pair for transmultiplexing of symbols without any SI. The merit of the proposed method to design a SA modifier matrix (SAM) for PAPR reduction is shown through performance comparisons for the application scenarios presented in this study.

Human face is a display of mental states that reflect the true feelings of a person. In this paper, we propose a framework for the video analysis of spontaneous facial expressions using an automatic facial emotion recognition system. Regional Hidden Markov Models (RHMMs) are created to describe the states of facial attributes for eyebrows, eyes, and mouth regions registered in a video sequence. The performance results reported in the paper show that the proposed technique outperforms the designated HMM for each emotion type [1, 2] tested with the Cohn-Kanade database for the person-independent case. More importantly, we used the proposed system to infer the mental states of a person based on spontaneous facial expressions. Merit of the proposed system is validated with human based evaluations.

The derivation of explicit Karhunen-Loève transform (KLT) kernel for an auto-regressive order one, AR(1), process has been of a major implementation concern. The main reason is due to difficulties in finding the roots of a transcendental tangent equation as required in the derivation. We propose a novel method to derive explicit KLT kernel for discrete AR(1) process. The merit of the proposed technique is highlighted. The proposed method has a good potential to make real-time KLT applications more feasible in the coming years.

The empirical correlation matrix of asset returns in an investment portfolio has its built-in noise due to market microstructure. This noise is usually eigenfiltered for robust risk analysis and management. Jacobi algorithm (JA) has been a popular eigensolver method due to its stability and efficient implementations. We present a fast FPGA implementation of parallel JA for noise filtering of empirical correlation matrix. Proposed FPGA implementation is compared with CPU and GPU implementations. It is shown that FPGA implementation of eigenfiltering operator in real-time significantly outperforms the others. We expect to see such emerging high performance DSP technologies to be widely used by the financial sector for real-time risk management and other tasks in the coming years.

We introduce the implementation of Generalized Discrete Fourier Transform (GDFT) with nonlinear phase on a Field Programmable Gate Array (FPGA.) After briefly revisiting the GDFT framework, we apply the framework to a channel equalization problem in an Orthogonal Frequency Division Multiplexing (OFDM) communication system. The block diagram of the system is introduced and detailed explanations of the implementation for each block are given along with the necessary VHDL code snippets. The resource usage and registered performance of the design is reported and alternatives to improve the design in terms of performance and resolution are provided. To the best of our knowledge, this is the first hardware implementation of GDFT reported in the literature.

We present fundamental concepts of risk and propose two methods for risk management of a portfolio in this paper. Moreover, we introduce their novel extensions to trading in multiple frequencies. We use stocks listed in NASDAQ 100 index as the investment universe for our back-testing to highlight the merit of the proposed portfolio risk management methods.

In this paper, we present design methods to optimize nonlinear phase functions of Generalized DFT (GDFT) for minimized auto and cross correlations. It is shown that GDFT offers sizable correlation improvements over DFT, Walsh, and Gold codes. We conclude the paper with a multiuser communications scenario where correlation optimized GDFT outperforms DFT and other known constant modulus codes in BER metric.

In this letter, the eigen-analysis-based optimal design method of orthogonal block transform, the so-called Karhunen-Loeve transform, is used to generate a new family of spread spectrum codes. These spread spectrum and varying power orthogonal codes might find their application in direct sequence code division multiple access (CDMA) communications. The basis functions of the proposed transform are jointly spread in the time and frequency domains. This property is in contrast with the traditional optimal block transforms maximizing energy compaction due to their frequency localized basis magnitude functions desired in source coding and spectral processing applications. Bit-error rate (BER) performance of the proposed CDMA spreading codes under both additive white Gaussian noise (AWGN) and Rayleigh flat fading channel conditions are compared with the industry standard codes like Walsh and Gold families. It is shown in this letter that the proposed technique allows us to design orthogonal spreading codes with flexible lengths that perform comparable to or better than the Gold and Walsh codes.

Walsh codes are perfectly orthogonal binary (antipodal) block codes that found many popular applications over several decades including synchronous multicarrier communications. It is well known that they perform poorly for asynchronous multiuser communications. Therefore, the Gold codes with their superior performance are the preferred user codes in asynchronous CDMA communications. In this paper, new sets of binary orthogonal and near orthogonal user codes are searched in the binary sample space and introduced for asynchronous spread spectrum multicarrier communications. It is shown that the proposed binary user code family outperforms the Walsh codes significantly and they match in performance with the popular, nearly orthogonal Gold codes closely for asynchronous multiuser communications in AWGN channels. On the other hand, we also observed that all the binary codes considered in this study performed comparable for Rayleigh flat fading channels except the inferior performance of Gold codes in synchronous case with high SNR. We present that there are a number of such new binary code sets with different lengths available in the binary sample space. They might help us to improve the multicarrier performance of future communications systems.

We analyze quantization based embedding and detection schemes in terms of the data hiding framework we have introduced (Ref.1, 2). This framework enables a better connection between analytical results and practical designs. We lay out the performance evaluation criteria and present comparison results for scalar quantization based data hiding techniques. Embedder-detector designs are evaluated based on three key issues. These are: (1) the type of post-processing employed at the embedder; (2) the form of demodulation; (3) the criteria used to optimize embedding-detection parameters. Data hiding methods are compared based on rate, correlation, and probability of error performance merits.

We explore the problem of proving ownership or origin of multimedia content like image/video or audio signals through watermarking. The need for watermarking arises out of the insufficiency of present copyright laws for claiming ownership of digital content. Watermarking schemes however, are threatened by counterfeit attacks, which primarily use the freedom available in choice of signature or choice of the watermarking method. A restrictive protocol for watermarking could go a long way in rendering counterfeit attacks extremely difficult. We suggest a comprehensive protocol as an extension of the one suggested by Craver et al. , that makes it possible for the true owner to claim ownership unambiguously, while making it extremely difficult for a pirate to do so.

The problem of efficient data hiding robust to lossy compression is divided into two subproblems. First is to maximize the resource-which is the permitted distortion of images. The second is the efficient use of the resource by means of sophisticated signaling techniques . We conclude that a good solution to the first problem is choosing magnitude discrete Fourier transform as the domain in which the message signal is embedded. For the signaling method we use a sophisticated signaling technique, using periodic functions to tile the space with signal constellations.

In this paper, we analyze oblivious (blind) information hiding methods from a new perspective. In 1983, Costa introduced a communications framework that also applies to oblivious information hiding. We present an alternate and equivalent framework by carrying out the channel dependent nature of the optimal encoder in a different manner. Within the proposed framework, decoder structure is simplified, although in effect it's a slender advantage compared to overall complexity. This interpretation provides a better connection between the analytical results and practical designs. We evaluate the practical embedding-detection schemes employing scalar quantization procedures along with thresholding and distortion compensation types of processings from this perspective. Furthermore, we justify the assumptions for the optimality of the two types of processings.

We analyze the system performance of code-division multiple-access (CDMA) random access systems with linear receivers and packet combing in multipath fading channels. Both slotted and unslotted CDMA systems with random spreading codes are considered. The analysis is based on large systems in which both the offered load and the processing gain tend to infinity but their ratio is fixed. It is relatively easy to characterize the traffic in such large systems, which enables us to derive the system throughput and average delay. From the analysis results, it is observed that multiuser detection and packet combining substantially improve the system performance.

In a mature watermarking and steganography application, the security should lie in the unknow random keys not in the embedding algorithms. In this paper, the problem of the selection of watermarking random sequences is investigated. In the widely used Spread Spectrum modulation schemes, a white sequence and a Low-Pass (LP) type sequences are analyzed, including its security level and synchronization requirements. It is found that the white independent is more secure but not robust against LP attack. Also it is not energy efficient. Actually The random sequence selection is usually the tradeoff between these factors. To be resilient to the removal attacks, the watermark spectrum should be similar to the spectrum of the cover signal. A watermarking sequence with fixed amplitude spectrum shape and random phase is proposed in the last part of the paper and its properties are discussed in detail.

Compression is the most common application that an forms of multimedia data undergo. Lossy nature of the compression due to the use of quantizers should be taken into account by the watermarking technique employed. In this paper, we present a data hiding scheme that incorporates the embedding with the quantization of lossy compression. We show that embedder-detector sets making use of compression scheme's quantization characteristics have better payload and lower compression bit rates. We also optimized the system parameters that maximize the hiding rate at different compression levels.

We investigate and prove the relationships among several commonly used and new performance metrics for embedded image bit streams transmitted over noisy channels. The average of the first error-free run length (FEFRL) is proposed as a simpler performance metric. On binary symmetric channels (BSC), the average FEFRL is obtained in closed form, which greatly simplifies performance optimization. Simulation results justify the merit of the proposed technique.

Robustness to compression is a basic requirement for any data hiding scheme. We concentrate on MP3 resistant oblivious data hiding. First the MP3 compression algorithm is briefly introduced. We then propose three effective data hiding schemes, where the message is embedded in amplitude, DFT phase domain, and noisy components, respectively. All these schemes demonstrate robustness to MP3 compression.

In oblivious steganography, host noise suppression is a great concern as the watermark signal energy is usually much less than that of the host signal. We model the data hiding as a more general H0/H1 hypothesis testing problem. Decision making is based on the statistical distinction between H0 and H1. A simple embedding scheme, viz., set partitioning, is proposed. The coefficients are divided into two sets to represent bit value 1 and 0. The average distortion introduced is calculated. Its optimum and suboptimal detection is discussed in detail. Analysis and simulation studies show improvement over existing schemes.

The performance characteristics of a multicode CDMA downlink with systematic convolutional (turbo) coding are obtained under a shadowed multipath fading mobile channel. Two different iterative decoding methods are compared in two different synchronous downlink schemes operating under given channel conditions with the same transmission bandwidth as in IS-95. Multicode CDMA with logMAP decoding is observed to provide better QoS with lower bit and frame error rates, reasonable complexity and shorter decoding delay compared to DS-CDMA downlink with single code pair assigned to each user.

This paper analyzes the performance of unslotted CDMA random access schemes with heavy-tailed packet length distribution. Most of the previous work cover the analysis of the system with Poisson process and for infinite population and infinite buffer size where the results mainly depend on the mean values of the traffic. We remove these assumptions and analyze the behavior of the system for the more general and realistic case of finite population, finite buffers and variable packet length. The emphasis is placed on the study of the effect of the heavy-tailed packet length on the system characteristics. Specifically, we show via analysis and simulation that even when the mean packet length is small the system can have bottlenecks due to the packet length characteristics, and therefore dynamic adaptation and control of the system is required.

This paper investigates parallel decoding of product code on the level of constituent decoders. Product code, based on a simple and short linear block code, allows the construction of fairly long codes with low complexity as well as a simple iterative decoding algorithm. We find that parallel decoding offers slight BER improvement over traditional serial decoding for three dimensional codes. More significantly, the results imply that parallelism could reduce the average decoding iterations.

In this paper, we present an information-theoretic approach to obtain an estimate of the number of bits that can be hidden in still images, or, the capacity of the data-hiding channel.We show how the addition of the message signal or signature in a suitable transform domain rather than the spatial domain can significantly increase the channel capacity. Most of the state-of-the-art schemes developed thus far for data-hiding have embedded bits in some transform domain, as it has always been implicitly understood that a decomposition would help. Though most methods reported in the literature use DCT or wavelet decomposition for data embedding, the choice of the transform is not obvious.We compare the achievable data-hiding capacities for different decompositions like DCT, DFT, Hadamard, and subband transforms and show that the magnitude DFT decomposition performs best among the ones compared.

We present a new method for optimal rate allocation between a progressive image coder and a channel coder for noisy channels. A mathematical model for the embedded bit streams is developed and used as a new metric for such a joint source-channel coding optimization problem. It is further shown that maximization of such a new metric is equivalent to the maximization of a conventional PSNR measure. PSNR improvements up to 0.3 dB over very noisy channels at low bit rate are achieved by using the proposed method without any additional overheads. Furthermore, the PSNR performance obtained by using the proposed method is upper bounded by those of the conventional Monte Carlo method. Computational results and numerical comparisons consistently verify the merit of the proposed technique.

The different QoS and data rate requirements of communication channels can be satisfied by varying one or more of the coding, digital modulation and spreading blocks. Keeping the transmission bandwidth fixed as in IS-95, four different interblock data rate configurations are compared. User powers are kept constant in the performance simulations. Various coding, modulation and spreading rate combinations yield different performance characteristics in AWGN, flat fading and multipath environments permitting trade-off decisions on choices of rate, QoS, and implementation.

A class of audio-visual content is segmented into dialogue scenes using the state transitions of a novel hidden Markov model (HMM). Each shot is classified using both the audio track and the visual content to determine the state/scene transitions of the model. After simulations with circular and left-to-right HMM topologies, it is observed that both performing very well with multi-modal inputs. Moreover, for the circular topology, the comparisons between different training and observation sets show that audio and face information together gives the most consistent results among different observation sets.

The minimum variance receiver has been previously proposed to suppress the multiple-access interference (MAI) blindly in CDMA channels by minimizing the receiver's output variance based on constrained optimization. In this paper, we propose a low-complexity minimum variance receiver that achieves almost the same performance as the high-complexity one with the optimal constraint vector. The proposed receiver is a RAKE receiver followed by an equal gain combiner. The impulse response of the filter in each finger is obtained by the minimum variance criterion. The low-rank minimum variance receiver uses a reduced-rank approximation of the estimated data covariance matrix. Its performance is better than the full-rank version of the receiver. The receiver is implemented adaptively with the same structure as a generalized sidelobe canceler.

The goal of this study is to investigate the performance of the watermarking scheme proposed by Ramkumar and Akansu (see IEEE Trans. on Multimedia, 1999) for watermarked images distorted by geometrical attacks such as StirMark. Although the geometric distortions are not visually noticeable, the watermark cannot be detected due to the resulting high mean square error between the original and attacked images. In order to extract the watermark from distorted images, the effects of these attacks must be minimized. This work focuses on the recovery of an attacked image using a reliable distortion estimation. The performance test is studied for several images attacked by StirMark. The proposed algorithm is shown to reduce false alarm probability in detection of the watermark from about 0.005 for (for StirMarked images) to less than 10/sup -50/.

This paper addresses good choices for the signaling method for multimedia steganography, or data hiding. A novel FFT based signaling method, with many properties that make it especially suitable for steganographic applications is proposed. The use of FFT permits large codebook sizes, without drastically increasing the computational complexity.

We present a dynamic unequal error protection (UEP) framework for the embedded image bit streams over noisy channels. A general source model for M-rate UEP schemes is derived and an optimization algorithm for 2-rate dynamic UEP schemes is developed. To facilitate the design and implementation of UEP schemes, we also derived a necessary condition and an upper bound for UEP gains. Simulation results demonstrate about 0.3 dB improvements over equal error protection (EEP) schemes at the price of negligible overheads while the progressiveness of the original bit streams is still kept.

It has been shown that cyclostationarity in the received signal allows the receiver to blindly identify the channel impulse response using only second-order statistics. In orthogonal frequency-division multiplexing (OFDM) systems, cyclostationarity is embedded at the transmitter due to cyclic prefix. In this paper, a subspace approach based on second-order statistics is proposed for blind channel identification in OFDM systems. We derive a sufficient condition that guarantees all the channels to be identifiable no matter what their zero locations are. Computer simulations demonstrate the superior performance of the proposed algorithm over methods reported earlier in the literature.

The minimum variance receiver has been proposed to blindly suppress the multiple-access interference in CDMA channels by minimizing the receiver's output variance based on constrained optimization. We consider the reduced-rank minimum variance receiver in asynchronous CDMA channels. The reduced-rank detector uses a reduced-rank approximation of the covariance matrix of the received data. When the covariance is known, the reduced-rank receiver cannot perform better than the full-rank receiver. However, when the covariance matrix is estimated, we demonstrate that the optimal reduced-rank receiver outperforms the full-rank receiver. The reduced-rank receiver is implemented adaptively using a subspace tracking algorithm; its performance is better than the conventional full-rank recursive least-square adaptive receiver.

Conventional spread spectrum communication methods essentially map a discrete symbol from an alphabet to a length-N real valued sequence, which represents a point in the N-dimensional signal constellation. The real valued sequence is in turn transmitted over a communications channel. However, communication schemes for the purpose of multimedia steganography (or data hiding) have to transmit the real valued sequence corresponding to a point in the signal constellation superimposed on the original or cover content (without affecting the fidelity of the original content noticeably). The detector therefore, has to estimate the origin of the signal constellation, and thereafter proceed to decode the transmitted symbol. In this paper, we explore some efficient solutions for signaling methods employing floating signal constellations.

A novel method is proposed to achieve efficient MPEG-2 video editing in compressed domain while preserving video buffer verifier (VBV) requirements. Different cases are determined, according to the VBV modes of the bit-streams to be concatenated. For each case, the minimum number of zero-stuffing bits or shortest waiting time between two streams is determined analytically, so that the resulting bit-stream is still VBV-compliant. The simulation results show that the proposed method is applicable to any MPEG-2 bit-stream independent of its encoder.

A robust watermarking scheme for images is proposed. The proposed method can be also be easily extended to other signals like audio and video. In the proposed method an optimal solution is obtained for maximizing the detection statistic (which is an indication of the degree of certainty with which the signature or the watermark is detected), for a given permitted distortion of the host signal, and additive noise variance. Other issues for improving the security of the watermarking scheme, like key based transforms, are also addressed.

We present an optimization algorithm for 3-rate dynamic unequal error protection (UEP) schemes for embedded image bit streams transmitted over noisy channels. A theoretical upper bound for UEP gains is also given. Simulation results demonstrate about 0.3 dB improvements over the optimal equal error protection (EEP) schemes at the price of negligible overheads and off-line computation load. The progression of the source bit streams is still kept. Furthermore, it is discovered that for binary symmetric channels (BSCs) with bit error rate (BER) satisfying /spl epsiv//spl les/10/sup -1/, at most 3 protection levels are enough to squeeze out almost all the UEP gains.

Channel reservation is a common solution to the call dropping problem in wireless cellular networks. We propose a predictive channel reservation scheme that is based on the mobility of the users. First, the concept of an influence curve is introduced to capture the influence that an active call exerts on the adjacent cells. Then, the amount of channels needed to be reserved in the adjacent cells can be adjusted dynamically according to the influence curve of each active call in the current cell. Finally, a call admission control scheme is proposed. Simulation results show that our channel reservation scheme outperforms the fixed reservation schemes.

A derivative of the set partitioning into hierarchical trees (SPIHT) image coding method, which generates substreams with different error-resilience properties, is proposed. By dividing the image bit stream into three classes, substreams with different immunity properties are obtained. The unequal protection of these substreams with different channel coding rates improves the overall performance of the method against channel errors. Simulation results show the superiority of the proposed method over some of the state-of-the-art methods.

Weyl-Heisenberg frames are complete signal representations corresponding to rectangular tiling of the time-frequency plane. Extensions of these frames are obtained in the rotated time-frequency planes by using the fractional Fourier transformation. It is shown that rotation does not affect the frame bounds. For some specific angles, lattices in rotated coordinates will map to the lattices in the Cartesian coordinates. The rotated Weyl-Heisenberg frames obtained are more suitable for chirp-like signal analysis and synthesis.

We present an information-theoretic approach to obtain an estimate of the number of bits that can be hidden in compressed image sequences. We show how addition of the message signal in a suitable transform domain rather than the spatial domain can significantly increase the data hiding capacity. We compare the data hiding capacities achievable with different block transforms and show that the choice of the transform should depend on the robustness needed. While it is better to choose transforms with good energy compaction property (like DCT, wavelet etc.) when the robustness required is low, transforms with poorer energy compaction property (like the Hadamard or Hartley transform) are preferable choices for higher robustness requirements.

We investigate the performance bounds for interference adaptive dynamic channel assignment (IA-DCA) in wireless networks. IA-DCA has been shown to be a promising radio resource allocation strategy in reuse systems such as FDMA and TDMA. Our previous work has presented the upper performance bound for IA-DCA algorithms by assuming an idealized dynamic channel allocation scheme with instant interference adaptation and maximal packing (IAMP) which not only provided guaranteed quality of received signal but also minimum co-channel reuse distance. The two sets of performance bounds previously obtained are results of exact analysis of the asymptotic case where the number of channels is assumed to be infinite, and an approximate analysis corresponding to the practical case where the number of channels is finite. In this paper, we study the practical case and verify the bounds by means of computer analysis. The idealized IAMP method is modified such that it reaches reasonable accuracy within finite simulation time.

In this paper, we present a novel design of a large-scale multicast ATM switch. We first propose a basic switch architecture, where switch inputs and outputs are grouped into small modules called input shared blocks (ISB) and output shared blocks (OSB). Link sharing with round robin (RR) cell scheduling solves output contention and eliminates the speedup requirement for the central switch fabric. To support non-uniform traffic, an enhanced switch architecture with a round robin prioritized output link reservation algorithm (RR+POLR) is further introduced. Both RR and RR+POLR algorithms use a group mapping scheme and distribute input cells' contention resolution to every ISB. Scheduling complexity is dramatically reduced. Simulations for a 256/spl times/256 switch demonstrate an improved performance under non-uniform traffic.

We present a data hiding scheme for still images, in which only the magnitude of the DFT coefficients are altered to embed the hidden information bits. We begin by identifying various components of a typical data hiding scheme, like the decomposition used, signature design, self-noise suppression and signaling. The final choice of components for the proposed data hiding scheme are well tailored by utilizing theoretical reasonings and experimental observations. Performance results show good robustness of the data hiding scheme for both JPEG and SPIHT image compression.

Multicarrier modulation techniques have been proposed for digital subscriber line (DSL) applications. The performance of DMT (OFDM) and DWMT (DSBMT) techniques for single and multitone interference are investigated. It is shown that a DMT system is sensitive to the location of narrow band interference. The DMT technique needs additional narrow band interference canceller before the forward FFT transform for performance improvements. In the DSBMT technique, due to a limited spectral overlap between its subcarriers, single (multi)-tone interference could affect only a few subchannels which correspond to these interferences. The DSBMT has a superior performance than DMT.

A novel time-frequency exciser is developed for the removal of chirp-like interferences in direct sequence spread spectrum (DSSS) communications. The chirplet decomposition iteratively expands signals in terms of time-frequency localized waveforms. The interference signal components are highly correlated with chirplets, and are represented by a few of the highest energy components of the decomposition. These components are excised from the received signal, and an excised signal goes through a detector for a decision. The proposed time-frequency exciser outperforms the conventional Fourier transform based excisers for chirp-like interference classes.

A novel nonmaximally decimated multirate filterbank structure is proposed for blind identification of communication channels. This structure is shown to be very similar to a form proposed earlier in the literature. The proposed blind channel identification algorithm is not sensitive to the characteristics of the unknown channel, including mixed phase and zeros on the unit circle. An optimal minimum mean square error based linear equalizer using the blind channel identification scheme is investigated. It is shown that the proposed system outperforms the existing zero-forcing blind equalization algorithms in the literature. It can simultaneously cancel the intersymbol interference (ISI) and suppress the noise enhancement. The reconstructed signal to noise ratio is maximized by the proposed algorithm. Simulation results show the superior performance and robustness of the proposed blind identification and equalization scheme.

We evaluate the performance of the copy network under three scenarios: (1) non-buffer non-splitting copy network (NBNS), (2) shared-input buffer-non-splitting copy network (SIBNS), and (3) shared-input-buffer-splitting copy network (SIBS). For NBNS, we derived the exact overflow and cell loss probabilities instead of the Chernoff bound. Furthermore, we propose a general Markov model, a novel theoretical approach, for the performance analysis of the shared-input-buffer copy networks. This analysis method can be applied for both SIBNS and SIBS. In our work, theoretical and simulation results are presented for every scenario. The comparison shows that improved performance of lower cell loss and higher throughput can be achieved by shared-input-buffer (SIBNS and SIBS). However the tradeoff is the long cell delay. With the splitting algorithm, SIBS can provide better performance than the NBNS and SIBNS.

We present an information-theoretic approach to obtain an estimate of the number of bits that can be hidden in still images, or, the capacity of the data-hiding channel. We show how addition of the message signal in a suitable transform domain rather than the spatial domain can significantly increase the channel capacity. We compare the capacities achievable with different decompositions like DCT, DFT, Hadamard, and subband transforms.

Interference adaptive dynamic channel allocation (IA-DCA) schemes are promising channel assignment strategies in TDMA/FDMA systems. The performance analysis of an IA-DCA scheme focuses on three factors: co-channel interference, traffic distribution, and the processing power of the DCA algorithm. In our previous work the impact of co-channel interference was studied. This paper investigates the impact of the traffic load distribution. As far as the processing power is concerned, an idealized interference adaptive maximum packing scheme is assumed such that it is able to deal with the traffic fluctuation among cells. This scheme gives the best performance for all IA-DCA algorithms under some assumptions. The combined effect of all three factors are considered in deriving an upper performance bound for IA-DCA algorithms in TDMA/FDMA systems.

Subband coding has been successfully used for image and video applications. This paper emphasizes two design issues of a subband codec. Namely, the selection of the filterbank family and the width of zero-zones in zero-zone linear quantizers. It is shown that the shift only filters (multiplierless PR-QMF) performs as good as other popular filterbank families. Finally, the significance of the quantizers' zerozones in the codec's rate-distortion performance is presented. It is found that the proposed efficient subband image/video codec performs comparable with the leading techniques like embedded zero-tree wavelet (EZW) coding reported in the literature. We also incorporated this efficient subband codec into a motion compensated video coding algorithm. It is observed that the superior performance of zero-zone linear quantizers is also valid for subband video coding. The subjective performance of the proposed efficient codec in image and video coding are found consistent with the objective performance results obtained.

An orthogonal transmultiplexer (synthesis/analysis filter bank) utilizing DFT (discrete fourier transform) basis is the current DMT (discrete multitone) standard for HDSL communications. The measures of the ICI and ISI distortions in a multicarrier modulation scheme are presented. The ICI and ISI performance of DFT along with other known linear transforms are compared. The issues of optimal basis design for a given channel are discussed. The available design tools of the subband transform theory are highlighted for the future performance improvements.

In this paper we propose a simple fast-search VQ of the LSF's to be used on top of the split VQ, that is, in each of the sub-vector domains.The main trait of the proposed method is that no suboptimal codebooks are used and there is no further reduction in performance.In each sub-vector domain, a full-size optimally trained codebook, typically of size 1024, is searched using a fastsearch algorithm.The result of this search is identical to that of a full-search, yet, only about 25% of a full-search complexity is needed.

This correspondence deals with suboptimal multiplierless perfect reconstruction quadrature mirror filter (PR-QMF) solutions. It is shown that multiplierless PR-QMFs perform comparable to or better than the known filter banks and discrete-cosine-transform-based (DCT-based) image coding techniques objectively and subjectively. They are very efficient to implement on very large scale integrated (VLSI) systems. These PR-QMFs might find uses in real-time image and video coding and other applications.

A smart adaptive time-frequency (ATF) excision algorithm is proposed to perform interference suppression in spread spectrum communications. The ATF exciser has the intelligence of deciding the domain of the excision. Additionally, adaptive subband transforms are utilized for frequency domain excision in order to track the spectral variations of the incoming signal. This adaptive transform approach brings significant performance improvements in spectral resolution. Bit error rate (EER) performance of the novel ATF exciser-based direct sequence spread spectrum (DSSS) system receiver is compared with other existing transform domain-based DSSS receivers. Time localized or frequency-localized interferer types and additive white Gaussian noise (AWGN) channels are considered in the study. In all cases, the smart ATF exciser-based DSSS receiver drastically outperforms the existing competitors. Its robust performance against the variations of the interferers is verified.

A smart algorithm for the decomposition of a channel into its subchannel in the discrete multitone communications is proposed in this paper. This algorithm evaluates the unevenness and energy distribution of the channel spectrum in order to get a variable adaptive partitioning of the channel. It is shown that the proposed algorithm leads to a near optimum performance of the discrete multitone transceiver. This flexible and smart splitting of the channel suffers less from the aliasing problem that exists in blind decompositions using fixed transforms. This paper extends the fixed discrete multitone to the flexible unequal bandwidth multichannel concept which has significant potentials for performance improvements in high-speed digital communications.

This study considers the transmission of digital signals in-band on-channel for application in digital audio broadcasting (DAB). In the FM-channel used for analog transmission, two different DAB scenarios are simulated. In the first approach, the digital data is spread in form of a weak direct sequence spread spectrum (DSSS) signal underneath the FM. Since the FM signal is a multitone interference to the DSSS signal, it is to be excised. In this paper, a new adaptive frequency domain technique is proposed for excising multitone interference in DAB, The second scenario performs an amplitude modulation on the FM signal. This technique assures no interference in-between the new, digital, and the already existing analog service,.

The concept of progressive optimization is proposed for the design of hierarchical subband transforms. The time and frequency properties of the product filters in subband trees are discussed and evaluated. It is shown that the performance improvements in image coding are possible by using different PR-QMF banks at different nodes of a subband tree.>

This paper compares the objective and subjective performance of direct and hierarchical subband decomposition schemes for image coding. It is observed that the hierarchical subband schemes outperform the direct structures in image coding. It is also shown that the dyadic tree subband codec outperforms the full-tree case for the image coding experiments performed in this study.

An analysis of band energy distributions in perfect reconstruction (PR) multirate systems is presented, and the effects of aliasing are evaluated. A performance measure called nonaliasing energy ratio (NER) is defined. The merit of the new measure is emphasized with performance comparisons of the popular block transforms and two-band PR-QMF based hierarchical filter banks. It is shown that the new measure complements the widely used energy compaction measure and is consistent with the subjective performance results.>

The popular orthonormal decomposition techniques such as block transforms and filter banks for multiresolution signal representation are unified. Their comparative performance results based on signal energy compaction are presented for image and AR(1) sources. It is observed that the filter banks with computationally efficient filtering algorithms and irregular tree structures are potential competitors to the block transforms, particularly for image processing and coding applications.>

A low bit rate adaptive video coding technique is proposed. The redundancy within adjacent video frames is exploited by motion-compensated interframe prediction, and the motion-compensated frame difference (MCFD) signal is decomposed by using a quadrature mirror filter (QMF) structure. The subband signals are vector quantized adaptively. The MCFD signals are low correlated, hence transform coding techniques are not very suitable for this kind of signal source. At 0.25-0.3 bits/pixel range, 38-40-dB peak-to-peak signal-to-noise ratio is achieved with the proposed scheme for four monochrome test sequences used: cindy, mono, duo, and quartet. The visual quality of the encoded test sequences is good, almost indistinguishable from the originals. The proposed technique outperforms the adaptive transform coding and adaptive vector quantization techniques for the tested sequences.>

This paper describes a class of orthogonal binomial filters which provide a set of basis functions for a bank of perfect reconstruction Finite Impulse Response Quadrature Mirror Filters (FIR-QMF). These Binomial QMFs are shown to be the same filters as those derived from a discrete orthonormal wavelet approach by Daubechies [13]. The proposed filters can be implemented very efficiently with output scaling, but otherwise no multiply operations. The compaction performance of the proposed signal decomposition technique is computed and shown to be better than that of the DCT for the AR(1) signal models, and also for standard test images.

Theoretical performance analysis and test simulations showed that the DWHT (discrete Walsh-Hadamard transform) is as good as the DCT (discrete cosine transform) for the encoding of low-correlation MCFD (motion-compensated frame difference) signals. However, the straightforward algorithm for the DWHT makes it a better choice than the DCT. The transform vector quantization (TRVQ) technique is employed for low bit rates. It is also observed that the robustness of the VQ (vector quantization) technique for the MCFD signals is much better than that of the VQ of the still frame images.>