Deep neural networks (DNNs) are very efficient at object recognition but sometimes have an overestimated reliability for erroneous decisions. It is therefore very important to check if the distribution of images to be processed matches that of the training set. This has been studied as out-of-distribution (OoD) detection. This paper proposes an algorithm allowing classification as well as OoD detection, based on the generative classifier (GC) paradigm.While a discriminative classifier (DC) learns p(y|x) to achieve classification and a generative network learns p(x) to achieve OoD detection, the GC learns p(x|y) to achieve classification as well as OoD detection. Since generative models such as variational auto-encoders (VAEs) show good performance for unsupervised OoD detection, we propose to structure the latent variable of a VAE with a gaussian mixture model (GMM) while building a GC. The proposed model1is shown to outperform classical approaches based on DCs for far OoD and generative adversarial network (GAN) based models for near OoD without losing any performance for the classification task.

This paper presents a learning-based, lossless compression method for static point cloud geometry, based on context-adaptive arithmetic coding. Unlike most existing methods working in the octree domain, our encoder operates in a hybrid mode, mixing octree and voxel-based coding. We adaptively partition the point cloud into multi-resolution voxel blocks according to the point cloud structure, and use octree to signal the partitioning. On the one hand, octree representation can eliminate the sparsity in the point cloud. On the other hand, in the voxel domain, convolutions can be naturally expressed, and geometric information (i.e., planes, surfaces, etc.) is explicitly processed by a neural network. Our context model benefits from these properties and learns a probability distribution of the voxels using a deep convolutional neural network with masked filters, called VoxelDNN. Experiments show that our method outperforms the state-of-the-art MPEG G-PCC standard with average rate savings of 28% on a diverse set of point clouds from the Microsoft Voxelized Upper Bodies (MVUB) and MPEG. The implementation is available at https://github.com/Weafre/VoxelDNN.

This paper proposes a lossless point cloud (PC) geometry compression method that uses neural networks to estimate the probability distribution of voxel occupancy. First, to take into account the PC sparsity, our method adaptively partitions a point cloud into multiple voxel block sizes. This partitioning is signalled via an octree. Second, we employ a deep auto-regressive generative model to estimate the occupancy probability of each voxel given the previously encoded ones. We then employ the estimated probabilities to code efficiently a block using a context-based arithmetic coder. Our context has variable size and can expand beyond the current block to learn more accurate probabilities. We also consider using data augmentation techniques to increase the generalization capability of the learned probability models, in particular in the presence of noise and lower-density point clouds. Experimental evaluation, performed on a variety of point clouds from four different datasets and with diverse characteristics, demonstrates that our method reduces significantly (by up to 37%) the rate for lossless coding compared to the state-of-the-art MPEG codec.

This paper addresses the problem of removing from a set of data (here images) a given private information, while still allowing other utilities on the processed data. This is obtained by training concurrently a GAN-like discriminator and an autoencoder. The optimization of the resulting structure involves a novel surrogate of the misclassification probability of the information to remove. Several examples are given, demonstrating that a good level of privacy can be obtained on images at the cost of the introduction of very small artifacts.

In order to improve the transmission reliability in current wireless communication systems, the Hybrid Automatic ReQuest (HARQ) protocol is employed to manage the unknown time-varying channel. The acknowledgments are fed back with delay on the return link. To fill up the idle time between a transmission and its acknowledgment, parallel HARQ streams associated with different messages are carried out. In this paper we improve on parallel HARQ by proposing a multi-layer HARQ protocol (also called superposition coding or multi-packet HARQ), where a single transmission may carry information on multiple messages. The multi-layer HARQ protocol works in presence of delay on the return link as parallel HARQ does, and does not require additional feedback such as the channel state information. It aims at improving the accuracy as well as the user's delay distribution, thus achieving throughput increase. Assuming capacity-achieving codes, we show that the proposed protocol outperforms parallel HARQ in throughput, message error rate, and delay distribution. Using practical codes and decoding algorithms the gains are as well significant, at the expense of the receiver's complexity.

A transform coefficient coding scheme is proposed for 4 × 4 blocks in Versatile Video Coding (VVC), targeting screen content applications. The proposed algorithm, called Unary Bitplane Coding (UBC), uses unary codes of the coefficient amplitudes and represents each block by their bitplanes. This representation allows exploiting further contextual information for source separation during the entropy coding. Experiments in the Joint Exploration test Model (JEM) show that replacing the existing transform coding with UBC only for 4 × 4 blocks brings on average 2.8% and 3.4% BD-R gain in the random access and all intra modes, respectively.

An intra coding algorithm with layer separation is proposed. This algorithm is designed on top of an adopted tool in VVC, called Block DPCM (BDPCM), and benefits from texture information in a neighborhood to derive intensity levels of background and foreground layers. This information is used to reduce large rate of residual in case of incorrect layer prediction by BDPCM. For this purpose, three inter-layer transition states are defined that are either implicitly or explicitly conveyed to the decoder. Once a transition is signaled, the decoder corrects the prediction value using the derived layer information. Experiments on screen contents show a BD-rate gain of about 10% percent over VVC Test Model (VTM) and 1% over the regular BDPCM, with the cost of computational complexity.

This paper considers wireless communication protocols implementing both cooperation and hybrid automatic repeat request. Based on a simple example (one source, one relay, and one destination), we propose the use of probabilistic protocols as an alternative to classical, deterministic ones. Starting initially with the finite state Markov chain (FSMC) analysis of a deterministic protocol, the idea of probabilistic protocol comes as an association with the simplified FSMC of the deterministic protocol. This FSMC contains two parameters that can be optimized for finding the best performance. So far, probabilistic protocols have been proposed at higher layers of communication systems, while we consider here the physical and MAC layers. In our analysis, the relay works in the demodulate-and-forward mode, and we demonstrate that: 1) the FSMC analysis predicts accurately the performance of the actual system and 2) the performance of the probabilistic protocol (when the two parameters are carefully tuned) outperforms the performance of a reference deterministic protocol. The analysis is checked upon Monte Carlo simulations.

In current wireless communication systems, the feedback required by the Hybrid Automatic ReQuest (HARQ) mechanism is received with some delay at the transmitter side. To alleviate this issue, parallal Stop-and-Wait HARQ is usually employed. In this paper, we propose a multi-packet HARQ protocol (also called superposition coding or multi-layer HARQ) to improve the user's delay distribution and increase the throughput, without any additional feedback such as Channel State Information. The performance analysis, provided from an information-theoretic point-of-view, shows that the proposed protocol offers better delay distribution, higher throughput and lower message error rate compared to the conventional parallel Stop-and-Wait HARQ, at the expense of increased decoding complexity.

A novel intra prediction algorithm is proposed to improve the coding performance of screen content for the emerging Versatile Video Coding (VVC) standard. The algorithm, called in-loop residual coding with scalar quantization, employs in-block pixels as reference rather than the regular out-block ones. To this end, an additional in-loop residual signal is used to partially reconstruct the block at the pixel level, during the prediction. The proposed algorithm is essentially designed to target high detail textures, where deep block partitioning structure is required. Therefore, it is implemented to operate on 4× 4 blocks only, where further block split is not allowed and the standard algorithm is still unable to properly predict the texture. Experiments in the Joint Exploration Model (JEM) reference software show that the proposed algorithm brings a Bjontegaard Delta (BD)-rate gain of 13% on synthetic content, with a negligible computational complexity overhead at both encoder and decoder sides.

Consider a source coding problem in presence of two dependent with memory sources (X, Y), for which only X is available at the encoder (referred to Alice). We first study the design of vector quantization for the situation where one of the source outputs, i.e., X, must be transmitted to the receiver (referred to Bob) within a prescribed distortion tolerance as in ordinary source coding. On the other hand, the other source, i.e., Y, has to be kept as secret as possible from the receiver or wiretappers. We next consider the opposite case where Y represents a relevant utility sequence to be reconstructed at Bob while trying to keep information about X secret from an eventual eavesdropper. A practical application involving electric consumption data measured from real houses is finally investigated.

A few years after standardization of the High Efficiency Video Coding (HEVC), now the Joint Video Exploration Team (JVET) group is exploring post-HEVC video compression technologies. In the intra prediction domain, this effort has resulted in an algorithm with 67 internal modes, new filters and tools which significantly improve HEVC. However, the improved algorithm still suffers from the long distance prediction inaccuracy problem. In this paper, we propose an In-Loop Residual coding Intra Prediction (ILR-IP) algorithm which utilizes inner-block reconstructed pixels as references to reduce the distance from predicted pixels. This is done by using the ILR signal for partially reconstructing each pixel, right after its prediction and before its block-level out-loop residual calculation. The ILR signal is decided in the rate-distortion sense, by a brute-force search on a QP-dependent finite codebook that is known to the decoder. Experiments show that the proposed ILR-IP algorithm improves the existing method in the Joint Exploration Model (JEM) up to 0.45% in terms of bit rate saving, without complexity overhead at the decoder side.

Relaying protocols are usually deterministic, i.e. they commute from one action to the next one in a predefined manner, depending on the success of the transmission. This paper demonstrates the advantages that probabilistic protocols (i.e. the next action to be undertaken depends on some probability) may have in this context. The whole study is performed on a simple source-relay-destination network, the relay working in demodulate and forward mode, in order to obtain insights on the interaction between relaying and HARQ. So far, probabilistic protocols have been mainly proposed for higher layers of communication systems, but are applied here to physical and MAC layers. Since it contains only two parameters, we demonstrate that our probabilistic protocol can easily be tuned for best performance using a Finite State Markov Chain (FSMC), and that it brings improvement over deterministic protocols in the same scenario.

Transmit power allocation techniques are very important to manage interference in small-cell networks. While available power allocation algorithms in the literature rely on a predefined routing protocol, we propose in this paper a power-efficient two-step algorithm that allows power allocation and routing to be performed jointly in femto-relay networks. First, we propose an interference-based partitioning method to cluster the femto-relays, then we adopt an iterative and distributed algorithm, inspired from game theory, for efficient transmit power allocation. We show that the corresponding power allocation game possesses a pure Nash equilibrium which is reached by the proposed algorithm within a number of iterations per femto-relay which can be as small as 1. Moreover, we show that our approach grants significant improvements in terms of power consumption, and permits the total consumed power to be divided by about 6 and 3 when respectively compared to the direct transmission and shortest path techniques.

The design of scalar quantization for distributed binary decision in presence of an eavesdropper (Eve) is investigated. An encoder/quantizer (Alice) observes a memoryless source and communicate via a public noiseless rate-limited channel with the detector (Bob) who has also access to a correlated analog source. Bob can take advantage of both informations to perform a binary decision on the joint probability law of these observations. Eve is further assumed to have access to a different correlated analog source and perfectly observe the information bits sent by Alice. This paper evaluates the various tradeoffs between the probabilities of error (on the decision) depending on the amount of information leakage from Alice to Eve. The Bhattacharyya distance; one of the distances measuring the difference between two probability distributions; is taken as a criterion to optimize the scalar quantizer subject to a tolerable constraint on the information leakage at the level of Eve. Numerical results for memoryless Gaussian sources demonstrate the performance of the proposed quantization method.

This paper studies secure communication based on incremental redundancy (INR) secure hybrid automatic retransmission request (HARQ) protocol over block-fading wiretap channels. The transmitter has no instantaneous channel state information (CSI) available from either main channel or the eavesdropper channel, hence the coding rates cannot be adapted to instantaneous channel conditions. We investigate the outage performance for two schemes of INR secure HARQ protocols: case 1) when there exists two reliable multibit feedback channels from both legitimate receiver and eavesdropper to the transmitter carrying a function of outdated CSI, and case 2) when there is a multibit feedback channel only from legitimate receiver. In both cases, we demonstrate that using the information carried via multibit feedback channels, the transmitter can adapt the coding rates in order to achieve a better secrecy throughput using a smaller number of transmissions comparing to the ACK/NACK feedback channel model. For some parameters, our rate adaptation protocol achieves a strictly positive secrecy throughput whereas it is equal to zero for the protocol with ACK/NACK feedback. We show that for some set of parameters, the loss of secrecy throughput between case 1 and case 2 is very small compared to the gain provided by both protocols.

Network Coded Cooperation (NCC) is known to provide full diversity order and high spectral efficiency for uncoded cooperative networks. However, the understanding of NCC applied to signals that have been protected by some Forward Error Correction (FEC) codes is still limited. This letter analyzes the diversity order attainable by NCC with channel coding (Coded-NCC) in a network topology with multiple sources, one relay and in the presence of fast Rayleigh fading. Due to the difficulty of characterizing the exchange of information between the network decoder and the channel decoder, iterative network and channel decoding algorithms are usually studied with the aid of simulations. In this letter, we overcome this limitation by proposing a near-optimal receiver that performs network decoding and channel decoding in a single decoding step of an equivalent super code. An upper bound and a tight approximation of the Bit Error Rate (BER) for all sources are derived. Based on the upper bound, we analytically show that Coded-NCC achieves a diversity order equal to 2f, where f is the minimum distance of the FEC code. This result generalizes those available for cooperative networks in the absence of channel coding (Uncoded-NCC), where the diversity order is equal to 2, as well as those available for coded transmission but without cooperation, where the diversity order is equal to f.

This paper considers the Broadcast Channel with Confidential Message (BCCM) where the sender attempts to send altogether a common message to two receivers and a confidential message to one of them. The achievable rate regions are derived for the power-constrained Gaussian BCCM with finite input alphabet using various transmission strategies. Namely, time sharing, superposition modulation and superposition coding are used as broadcast strategies. For superposition modulation and superposition coding, the maximal achievable rate regions are obtained by maximizing over both constellation symbol positions and the joint probability distribution. The maximization of the secrecy rate for wiretap channels is also studied as a particular case of the BCCM problem. We compare the considered transmission strategies in terms of percentage gains in achievable rates. We concentrate on the impact of the finite alphabet constraint on achievable rates, and show that this constraint may change well known results obtained in the Gaussian case. We show also that the secrecy constraint can change the shape of the achievable rate region in superposition modulation used in some standards when symbols are equiprobable. On a more practical side, it is shown that a performance close to the optimum can be obtained by strategies with reduced complexity.

In this paper, we study outage probability (OP) and diversity order of a M-source and N-relay wireless network that combines network coding (NC) and relay selection (RS). More specifically, a decode-and-forward (DF) relaying protocol is considered and the network-encoding vectors at the relays are assumed to constitute a maximum distance separable (MDS) code. Single relay selection (SRS) and multiple relay selection (MRS) protocols are investigated, where the best relay and the L best relays forward the network-coded packets to the destination, respectively. An accurate mathematical framework for computing the OP is provided and from its direct inspection the following conclusions on the achievable diversity are drawn: 1) the SRS protocol achieves diversity order equal to two regardless of M and N and 2) the MRS protocol achieves diversity order equal to L + 1 if L <; MandequaltoN + 1 if L ≥ M.These analytical findings are substantiated with the aid of Monte Carlo simulations, which also show that RS provides a better OP than NC based on repetition coding if L ≥ M.

This paper investigates the maximization of the secrecy-achievable rate region for the Gaussian broadcast channel with confidential message (BCCM) using finite input constellations. The maximization is done jointly over symbol positions and their joint probabilities. The secrecy-achievable rate regions are given for various broadcast strategies which differ in their complexity of implementation. We compare these strategies in terms of improvement in achievable rates and we study the impact of finite input alphabet on the secrecy-achievable rates. It is shown that finite alphabet constraints may change well known results holding in the Gaussian case.

This paper deals with synchronization problem in IEEE 802.11a wireless system. In addition to traditional training sequences, the SIGNAL field of the physical frame can be considered as a new source of information. Indeed the receiver is able to predict the SIGNAL unknown parts relying on the knowledge provided by the CSMA/CA protocol during the negotiation of the transmission medium reservation. The exchanged RtS control frame used jointly with the bit-rate adaptation algorithm to the channel helps the receiver not only to predict the SIGNAL field but also to get information on the channel state. Based on this knowledge, joint MAP channel, time and frequency synchronization algorithm is carried out. Moreover to estimate the residual time offset, a timing metric in frequency domain is performed by minimizing the average of transmission errors in the presence of all channel estimation errors. The performance in terms of probability of synchronization failure is shown to be improved compared to existing algorithms.

This paper presents a joint protocol-channel decoding techniques to improve the estimation quality of the type of a packet corrupted by transmission errors. It exploits the redundancy present in the protocol stack to determine the packet type which is the most probable a posteriori. An optimal and a suboptimal estimation algorithm are presented. The latter is illustrated on the type estimation of packets compressed with the Robust Header Compression algorithm in unidirectional mode. Compared to a classical packet type estimation technique, improvements of more than 3.9 dB are observed in terms of channel SNR at header error rates of 10-2.1

This paper studies a Gaussian two-way relay channel where two communication nodes exchange messages with each other via a relay. It is assumed that all nodes operate in half-duplex mode without any direct link between the communication nodes. A compress-and-forward relaying strategy using nested lattice codes is first proposed. Then, the proposed scheme is improved by performing layered coding: A common layer is decoded by both receivers, and a refinement layer is recovered only by the receiver that has the best channel conditions. The achievable rates of the new scheme are characterized and are shown to be higher than those provided by the decode-and-forward strategy in some regions.

This paper addresses the time synchronization problem in IEEE 802.11a OFDM wireless systems. To enhance the coarse time synchronization mechanism, recent methods exploit not only traditional training sequences as specified by the standard but also additional knowledge available when the Carrier Sense Multiple Access with Collision Avoidance mechanism (CSMA/CA) is triggered. In this case, additional information can be used as training sequences based on the protocol knowledge training sequence known by the receiver. This step is followed by a time synchronization and channel estimation which results in the smallest Channel Estimate Errors (CEE) according to the selected criterion (e.g. LS, MAP). However it was found that the synchronization failure probability heavily depends on the channel estimate quality. Therefore to improve the performance of this class of algorithms, we propose an optimal time synchronization metric that minimizes the average of the transmission error over all CEE. Simulation results show a strongly improved performance in terms of synchronization failure probability in comparison with the existing algorithms.

Cross-layer error protection strategies couple systematic error correction at the physical layer with adaptive data retransmission at the data-link layer, to mitigate the effects of fast-fading propagation channels. This work, in the form of a case-study, aims to introduce useful tools for the assessment and analysis of the energetic performance of such strategies. In particular, it is shown that in energy-constrained applications, such as wireless sensor networks, the design guidelines should be based on a precise definition of the energy per bit. This definition also allows to evaluate the minimum energy that is necessary to transmit data in a given setting.

This paper studies the four-node Multiple Access Relay Channel (MARC) under quasi-static block Rayleigh fading channels and Gaussian noise environment. A relay employs Demodulate-and-Forward (DMF) protocol in order to help two channel-encoded sources to communicate with a destination. The contributions of the paper are threefold: i) we propose a Near Optimal Joint Network/Channel decoding (NO-JNCD) algorithm at the destination. The NO-JNCD employs Cooperative Maximum Ratio Combining (C-MRC) detector and the BCJR algorithm applied to a compound code which trellis consists of all possible states of two single trellises at sources; ii) we compute extended distance spectrum of the compound code containing input weights for each source and output weights on each fading channel; and iii) from the extended distance spectrum, we derive the Bit Error Rate (BER) upper bound for the compound code as well as for each source. It is shown by simulation that the proposed decoder provides performance very close to that of the optimal JNCD with both DMF and Decode-and-Forward (DF) relaying protocols. Finally, the analysis is checked by simulation.

Parallel two-way relay channel models a cooperative communication scenario where a relay helps two terminals to exchange their messages over independent Gaussian channels. For the single channel case, we have shown previously that lattice-based physical layer network coding achieves the same rate as compress-and-forward scheme with a random coding strategy. A direct extension of this lattice-based scheme to parallel Gaussian channel is to repeat the same strategy for each subchannel. However this approach is not scalable with the number of sub-channels since the complexity of the scheme becomes prohibitive when a large number of sub-channels is employed. In this contribution, we investigate a lattice-based physical layer network coding scheme where the relay jointly processes all the sub-channels together. We characterize the rate region allowed by our coding scheme and assess the performance penalty compared to the separate channel processing approach.

This paper considers incremental redundancy hybrid ARQ (HARQ) transmission over independent block-fading channels. The transmitter, having no knowledge of the instantaneous channel state information (CSI) can or - allocate the transmission rate knowing the statistics of the channel, or - adapt the transmissions rates using the outdated CSI, i.e., the one experienced by the receiver in the past transmissions that resulted in a packet decoding failure. Aiming at throughput maximization problems under constraint on the outage probability, we show how to optimize the rate-adaptation and rate-allocation policies using dynamic programming framework. Numerical examples obtained in a Rayleigh-fading channel show that rate adaptation provides notable gains over a rate allocation and non-adaptive HARQ, and, for high SNR, only a few transmissions are necessary to approach closely the ergodic capacity.

This paper studies the problem of finding the optimal route of a stream between the transmitter and the receiver in an ad hoc network while minimizing the total power consumption. The criterion is the minimum power, and the fact that the stream must reach its destination is used as a constraint. Compared to previous approaches, we introduce more flexibility by allowing the stream to be split between several intermediate destinations, and we do not use the specific approximations (mainly high SNR) required to make the problem convex. Therefore, the main difficulty is due to the non convexity of the problem. In this contribution, we address a simple situation (flat fading channels, constant during the optimization process, simple transmitters and receivers) in order to demonstrate some properties of the problem. Finally, we propose an algorithm which, at least in some cases, is able to overcome the non convexity. Illustrations of the path followed by the algorithm demonstrates that it is able to find the global optimum even is difficult cases.

In this paper, we study the three-node multiple-access relay channel under realistic operating conditions. More specifically, in our setup all wireless links are subjected to Rayleigh fading and Gaussian noise, and the relay node can be located at different distances from source and destination nodes. We assume that coded bits are fully-interleaved, so that the transmitted symbols are mutually independent. By exploiting factor graph representation, we propose a joint network/channel decoding algorithm, which takes into account decoding errors introduced by the relay node. The convergence of the proposed iterative decoding algorithm is studied through EXIT chart analysis. The proposed decoder is compared with a recently proposed algorithm, which employs the demodulate-and-forward protocol and works on coded bits, and it is shown that it provides better performance. More specifically, for a bit error rate (BER) equal to 1e−3, it is 1.5dB better when the relay is located near the destination and 2.0dB better when the relay is in between source and destination.

Channel decoding makes use of redundancy in the coded bits. However, any redundancy in the information bitstream can also improve the decoding process. This paper shows that a careful examination of communication standards (including the headers added by the network layers) exhibits some fields which can be considered as redundant, either based on already received packets, or on the standard. The efficient use of this knowledge in the channel decoding process is denoted as protocol-assisted channel decoding. Assuming perfect synchronization and available channel state information, the proposed method applied on 802.11a PHY and MAC layers provides a substantial link budget improvement without modifying the standard, while the introduction of an additional interleaver provides additional bit error rate improvements.

In many communication standards, several variable length frames generated by some source coder may be aggregated at a given layer of the protocol stack in the same burst to be transmitted. This decreases the signalization overhead and increases the throughput. However, after a transmission over a noisy channel, Frame Synchronization (FS), i.e., recovery of the aggregated frames, may become difficult due to errors affecting the bursts. This paper proposes several robust FS methods making use of the redundancy present in the protocol stack combined with channel soft information. A trellis-based FS algorithm is proposed first. Its efficiency is obtained at the cost of a large delay, since the whole burst must be available before beginning the processing, which might not be possible in some applications. Thus, a low-delay and reduced-complexity Sliding Window-based variant is introduced. Second, an improved version of an on-the-fly three-state automaton for FS is proposed. Bayesian hypothesis testing is performed to retrieve the correct FS. These methods are compared in the context of the WiMAX MAC layer when bursts are transmitted over Rayleigh fading channels.

This paper proposes an algorithm for the robust reception of compressed HTML files transmitted over a noisy mobile radio channel. Both source encoders and transmission systems are assumed to be standard compliant. The source encoder follows the HTTP1.1 protocol specifications, i.e. the HTML files are encoded by the deflate algorithm, a combination of Lempel-Ziv and Huffman algorithms. The transmission scheme follows IEEE 802.11a (and 802.11n) standard as an example. The proposed receiver is based on an iterative joint source-channel decoding approach. The Soft-Input Soft-Output outer source decoder is based on a sequential M-algorithm, which has been modified to improve the decoding performance by exploiting the specific grammatical and syntax rules of (i) Huffman codes; (ii) Lempel-Ziv codes; and (iii) HTML language. Simulation results following the IEEE 802.11a (and 802.11n) standard over additive white Gaussian noise and Rayleigh fading channels show that the proposed receiver drastically reduces the number of errors occurring in the received HTML files compared to the classical receivers. An EXIT chart analysis illustrates some properties of this combination of source and channel decoders.

We analyze the throughout achievable in transmission over block-fading channels when the instantaneous channel state information (CSI) is not available at the transmitter. Assuming operation with incremental redundancy hybrid ARQ (HARQ), we propose to adapt the transmissions rates using the outdated CSI, i.e., the one experienced by the receiver in the past transmissions that resulted in a packet decoding failure. We show that, even if the CSI is fully outdated, i.e., in independently block-fading channel, the adaptation provides notable gains over a non-adaptive HARQ and for high SNR, a few transmissions are necessary to approach closely the ergodic capacity.

Iterative decoding is considered in this paper from an optimization point of view. Starting from the optimal maximum likelihood decoding, a (tractable) approximate criterion is derived. The global maximum of the approximate criterion is analyzed: the maximum likelihood solution can be retrieved from the approximate criterion in some particular cases. The classical equations of turbo-decoders can be obtained as an instance of an hybrid Jacobi/Gauss-Seidel implementation of the iterative maximization for the tractable criterion. The extrinsics are a natural consequence of this implementation. In the simulation part, we show a practical application of these results.

In this paper, we investigate joint network and channel decoding algorithms for the multiple-access relay channel. We consider a realistic reference scenario with Rayleigh fading over all the wireless links, including the source-to-relay channels. Our contribution is twofold: i) first, we develop the quasi-optimal joint network and channel decoder by taking into account possible errors over the source-to-relay channels, and ii) second, we propose a low-complexity iterative joint network and channel decoding algorithm, which reduces the number of channel decoders with respect to state-of-the-art solutions. Our numerical results show that: i) in fully-interleaved Rayleigh fading channels, the proposed solution provides almost the same bit error probability as the quasi-optimal scheme but with a reduction in complexity of approximately the 65%, and ii) in Rayleigh block-fading channels, the proposed scheme yields almost the same bit error probability as state-of-the-art solutions but with a reduction in complexity of approximately the 30%.

Frame Synchronization (FS) is required in several communication standards in order to recover the individual frames that have been aggregated in a burst. This paper proposes a low-delay and reduced-complexity Sliding Trellis (ST)-based FS technique, compared to our previously proposed trellis-based FS method. Each burst is divided into overlapping windows in which FS is performed. Useful information is propagated from one window to the next. The proposed method makes use of soft information provided by the channel, but also of all sources of redundancy present in the protocol stack. An illustration of our ST-based approach for the WiMAX Media Access Control (MAC) layer is provided. When FS is performed on bursts transmitted over Rayleigh fading channel, the ST-based approach reduces the FS latency and complexity at the cost of a very small performance degradation compared to our full complexity trellis-based FS and outperforms state-of-the-art FS techniques.

This paper presents a bandwidth-efficient method for improved decoding of critical data generated by the MPEG-4 AAC audio coder when encoded bitstreams are transmitted over noisy channels. The critical data of each encoded frame is estimated using the redundancy due to the correlation between successive frame headers and using an optional CRC as an error-correcting code. Simulation results for an AWGN channel show a substantial link budget improvement (more than 5 dB) compared to a classical hard-decoding method. Once critical data are efficiently estimated, the work of previously proposed joint source-channel decoding techniques to recover the remaining parts of the MPEG4 AAC frames is significantly facilitated.

Prediction methods by template matching are often mentioned to improve video coding efficiency. They are based on a Markovian model to find the most similar patterns of texture in previously encoded information. These kinds of methods are more efficient than H.264/AVC intra prediction modes in many cases, such as complex texture coding. However, the template matching method is not always optimal as to the best predictor choice. This paper introduces the use of a sorted set of template matching predictors to propose different texture patterns. At the encoder side, the best predictor is selected by minimizing a rate-distortion metric and is transmitted to the decoder. Despite this needed additional side information, simulation results show that the proposed scheme further improves the coding efficiency of H.264/AVC up to 11%.

Video transmission on wireless links usually requires some frame aggregation, so that the overhead due to the headers is not a too large percentage of the bit-stream. In such a situation, header error detection is essential, because any error in the header may cause a loss of several consecutive frames. The contributions of this paper are (i) an improved reception of the length field of the header and (ii) an improved Frame Synchronization (FS) algorithm for aggregated frames. The FS algorithm has the following characteristics: (i) it makes use of the implicit redundancies which are present in the headers, thus resulting in an efficient synchronization of the variable-length frames; (ii) Bayesian hypothesis testing is used to estimate the correct synchronization; (iii) the proposed algorithm performs estimation on-the-fly and does not require reading the whole bit-stream. A comparative performance evaluation with respect to previously proposed algorithms is provided in the context of WiMAX MAC.

This paper presents an enhanced permeable layer mechanism useful for highly robust packetized multimedia transmission. Packet header recovery at various protocol layers using MAP estimation is the cornerstone of the proposed solution. The inherently available intra-layer and inter-layer header correlation proves to be very effective in selecting a reduced set of possible header configurations for further processing. The best candidate is then obtained through soft decoding of CRC protected data and CRC redundancy information itself. Simulation results for WiFi transmission using DBPSK modulated signals over AWGN channels show a substantial (4 to 12 dB) link budget improvement over classical hard decision procedures. We also introduce a sub-optimal and hardware realizable version of the proposed algorithm.

This paper presents a MAP estimation method allowing the robust decoding of compressed video streams by exploiting the bitstream structure (Le., information about the source, related to variable-length codes and source characteristics) together with the knowledge of the MAC layer CRC (here considered as additional redundancy on the MAC packet). This method is implemented via a sequential decoding algorithm in which the branch selection metric in the decoding trellis incorporates a CRC-dependent factor, and the paths which are not compatible with the source constraints are pruned. A first implementation of the proposed algorithm performs exact computations of the metrics, and is thus computationally expensive. Therefore, we also introduce a suboptimal (with tunable complexity) version of the proposed metric computation. This technique is then applied to the robust decoding of sequences encoded using the H.264/AVC standard based on CAVLC and transmitted using a WiFi-like packet structure. Significant link budget improvement results are demonstrated for BPSK modulated signals sent over AWGN channels, even in the presence of channel coding.

Iterative decoding was not originally introduced as the solution to an optimization problem rendering the analysis of its convergence very difficult. In this paper, we investigate the link between iterative decoding and classical optimization techniques. We first show that iterative decoding can be rephrased as two embedded minimization processes involving the Fermi-Dirac distance. Based on this new formulation, an hybrid proximal point algorithm is first derived with the additional advantage of decreasing a desired criterion. In a second part, an hybrid minimum entropy algorithm is proposed with improved performance compared to the classical iterative decoding. Even if this paper focus on iterative decoding for BICM, the results can be applied to the large class of turbo-like decoders.

The paper first recalls the Blahut Arimoto algorithm for computing the capacity of arbitrary discrete memoryless channels, as an example of an iterative algorithm working with probability density estimates. Then, a geometrical interpretation of this algorithm based on projections onto linear and exponential families of probabilities is provided. Finally, this understanding allows also to propose to write the Blahut-Arimoto algorithm, as a true proximal point algorithm. it is shown that the corresponding version has an improved convergence rate, compared to the initial algorithm, as well as in comparison with other improved versions.

Classical systems using hierarchical modulation (such as DVB-SH) involve a "high-priority" (HP) and a "low-priority" (LP) bit stream that are separately and independently encoded before being mapped on non-uniformly spaced constellation points, leading to different levels of error protection. However, an inherent drawback of this scheme is the severe performance degradation of the less protected LP stream. In this paper, we propose a concatenated encoding scheme that mixes the two streams in order to make the encoding of the LP stream dependent on the well protected HP stream. Consequently, at the receiver side, the reliable information on HP bits provided by the HP decoder is naturally involved in the decoding process of the LP stream and acts as "soft pilots". We have exploited this feature inside a soft decoding mechanism in order to improve the LP decoding performance while keeping the HP decoding performance unchanged. Numerical results presented in the case of the DVB-SH standard show that the proposed system outperforms the classical approach in terms of LP bit error rate without a considerable increase in the receiver complexity.

The objective of this paper is two fold: proposes a distributed resource optimization framework in the context of heterogeneous systems and optimizes allocation of discrete resources and bit rates to users within each system (network). In fact, most of existing wireless techniques use Adaptive Modulation and Coding technique where a finite set of Modulation and Coding schemes is available in the system. The bit rates to attribute to users are then selected among a finite set of available discrete rates. More specifically, the objective in this paper is to minimize to overall transmit power of the whole system while ensuring to each user a minimum data rate and fulfilling a maximum power constraint within each system. The use of centralized optimization frameworks to handle Radio Resource Management (RRM) problem in heterogeneous systems requires huge signaling overhead. In order to obtain distributed resource allocation, we approach the problem by using a convex optimization framework and thus split it into different subproblems each one performed by a given agent of the network(base station or user). We propose then a distributed algorithm based on successive iterations (negotiations) between the base stations and users. In each iteration, each network node (user or base station) has to handle a discrete optimization problem using an efficient (low complexity) algorithm described in the paper. By using in each iteration an appropriate update method, based on the subgradient of the dual Lagrangian objective function, our approach converges to the optimal solution with zero duality gap and after a finite number of iterations. Simulation results show good performance of the proposed algorithm and the convergence to the global optimal solution.

This paper presents an improved sequential MAP estimator to be used as a joint source-channel decoding technique for CABAC encoded data. The decoding process is compatible with realistic implementations of CABAC in standards like H.264, i.e, handling adaptive probabilities, context modeling and integer arithmetic coding. Soft-input decoding is obtained using an improved sequential decoding technique, which allows to obtain a tradeoff between complexity and efficiency. The algorithms are simulated in a context reminiscent of H264. Error detection is realized by exploiting on one side the properties of the binarization scheme and on the other side the redundancy left in the code string. As a result, the CABAC compression efficiency is preserved and no additional redundancy is introduced in the bit stream. Simulation results outline the efficiency of the proposed techniques for encoded data sent over AWGN and UMTS-OFDM channels.

This contribution provides a very high data rate UWB system design based on direct sequence spreading. In order to attain > 500Mbps for short range communication, two classical solutions are feasible: either use very short spreading sequences or very large bandwidth. These classical strategies increase intersymbol interference. In contrast, our approach consists in allocating multiple long codes to each user. Long codes allow to cope with channel impairments while the use of multiple codes maintains very high bit rates. With such high data rates the receiver obviously has to be simple by requiring a low complexity digital processing. This paper shows that the proposed strategy is compatible with simple receivers. In fact, even if the amount of interference increases, joint usage of very long codes and simple linear receiver can result in high data rate, high performance UWB system. This assertion is substantiated through comparisons with existing UWB solutions.

Joint source-channel decoding of variable length codes (VLC) for image and video transmission over wireless links is a subject of increasing interest. This paper proposes an optimum decoder of VLC sequences which exploits inherent redundancy in source data, namely the VLC syntax and source constraints. The proposed decoder is able to deliver both Maximum Likelihood optimum hard-output and soft-output solutions. The complexity of the algorithm is analyzed and a reduced complexity version of this algorithm with a slightly sub-optimal performance is proposed. An iterative joint source-channel decoding system is designed for demonstrating the performance of the proposed soft-input soft-output (SISO) decoder. Realistic contexts have been used for simulations: typical video sequences and typical radiomobile channels.

The optimal decoder achieving the outage capacity under imperfect channel estimation is investigated. First, by searching into the family of nearest neighbor decoders, which can be easily implemented on most practical coded modulation systems, we derive a decoding metric that minimizes the average of the transmission error probability over all channel estimation errors. Next, we specialize our general expression to obtain the corresponding decoding metric for fading MIMO channels. According to the notion of Estimation-induced outage (EIO) capacity introduced in our previous work and assuming a block Rayleigh-fading channel, we characterize the maximal achievable information rates using Gaussian codebooks associated to the proposed decoder. These achievable rates are compared to the rates achieved by the classical mismatched maximum likelihood (ML) decoder and the ultimate limits given by the EIO capacity. Numerical results show that the derived metric provides significant gains, in terms of achievable EIO rates and bit error rate (BER), in a bit interleaved coded modulation (BICM) framework, without introducing any additional decoding complexity. However, the achievable rates of such metric are still far from the EIO capacity.

Communication systems are usually designed by assuming perfect channel state information (CSI). However, in many practical scenarios, only a noisy estimate of the channel is available, which may strongly differ from the true channel. This imperfect CSI scenario is addressed by introducing the notion ofestimation-induced outage (EIO). We derive a single-letter characterization of the maximal EIO rate and prove an associated coding theorem and its strong converse for discrete memoryless channels (DMCs). The transmitter and the receiver rely on the channel estimate and the statistics of the estimate to construct codes that guarantee reliable communication with a certain outage probability. This ensures that in the non-outage case the transmission meets the target rate with small error probability, irrespective of the quality of the channel estimate. Applications of the EIO capacity to a single-antenna (nonergodic) Ricean fading channel are considered. The EIO capacity for this case is compared to the EIO rates of a communication system in which the receiver decodes by using a mismatched maximum-likelihood (ML) decoder. The effects of rate-limited feedback to provide the transmitter with quantized CSI are also investigated.

This paper concerns the error correction of corrupted compressed HTML pages during their transmission via a noisy mobile channel. The proposed receiver is based on an iterative joint source channel decoding approach similar to the turbo decoding of two serial concatenated codes. We propose a soft-input soft-output Lempel-Ziv inner decoder based on the modified version of the traditional sequential decoding M-algorithm. This new algorithm exploits the specific grammatical rules of the Lempel-Ziv-77 codes which are combined to the syntax of the HTML language. This source decoder is combined to a Soft-Input Soft-Output channel decoder of convolutional codes. Simulation results, over an additive white Gaussian noise channel, show that the proposed method drastically reduces the number of files in error compared to any conventional channel decoding.

This paper proposes a new still image codec. The encoder has the following structure: a set of pixels of the image is selected and transmitted, together with their position. Then, the value of the image at other places is obtained by a prediction algorithm at the decoder. A useful theoretical covariance model adapted to the image to be encoded is proposed, avoiding the transmission of additional information. The selected pixels and their corresponding positions on the image are encoded using lossless coding algorithms. The computational time of the decoding process is significantly reduced according to the efficient structured memory organization of (i) the image covariance values and (ii) the ordered distances concerning the search of nearest pixels. Experimental results performed on a set of test images show that the rate- distortion results are competitive to the best coders JPEG 2000 and SPIHT with arithmetic coding.

The design of efficient rateless coding schemes for multicast applications in wireless environments is investigated. First, the rateless paradigm for non-ergodic channels is introduced by making use of the dynamic-decoding nature of rateless codes that allows them to adapt opportunistically the code rate to the channel realization (assumed unknown at the transmitter). The information theoretical limits of such codes can be interpreted in terms of the notion of outage capacity. Then, we consider a quasi-static Rayleigh-fading channel with perfect and imperfect channel state information (CSI) at the receiver. We show that the optimal consistent measure of information for decoding with imperfect CSI, is given by the log-likelihood ratio (LLR) of the received bits via a composite (more noisy) channel. The optimization of Raptor codes, which depends on the delay requirements of decoding, is obtained by using Information content evolution under Gaussian approximation. Simulation results show that optimized Raptor codes can operate very close to the theoretical limits on a wide range of delay requirements.

This paper defines a retransmission scheme for transmitting video data which allows the use of robust decoding (such as joint source and channel decoding) jointly with acknowledgement procedures. Conventional video communications use CRC-based retransmissions to obtain a nominal quality of the video (defined by the source coder). A lot of work has also been done on robust video transmission, which aims at providing a video of the best possible quality, for a given channel. However, robust techniques are not compatible with CRC-based retransmissions, and the quality of the received signal cannot really be controlled. As a result, the throughput is the best possible one, at the cost of a reduced quality. The scheme we define allows the use of robust decoding together with retransmissions, thus improving the throughput of the system while trying to keep the nominal quality. As a result, the proposed scheme is compatible with most actual mechanisms. The method is based on an evaluation of the quality of the robust estimate, followed by an hypothesis testing to decide whether a retransmission is necessary or not. Simulations evaluate the quality and the average number of retransmissions.

We propose a new decoding algorithm based on a cascade structure for impulsive noise correction in multicarrier systems such as orthogonal frequency division multiplexing (OFDM) based systems. This new cascade structure based algorithm has the advantage to lead on one hand to a reduced complexity implementation in an OFDM demodulator and on the other hand to overcome the lacks of other algorithms dedicated for impulsive noise correction. To do so, we start the derivation by investigating closed-form analytical expressions of the a priori probabilities used in hypotheses tests that are necessary for the different steps needed for the execution of the cascade structure based decoding algorithm. Such closed-form expressions for the a priori probabilities, result in efficient setups of the threshold values used for the execution of the decoding algorithm. Furthermore, we show that this cascade structure can be applied to efficiently cancel the impulsive noise in OFDM systems. As the clipping event, present in OFDM systems, can be seen as an additive impulsive noise, we prove that after introducing some minor modifications, we can apply the proposed cascade structure to significantly reduce the peak to average power ratio (PAPR) level in such systems. Simulation results are given to illustrate the usefulness of the proposed algorithm results in practical OFDM based communication systems such as Hiperlan 2.

Semifragile watermarking can be used to detect illegal and local manipulation, while being robust toward legal processings such as lossy compression. The aim of this work is twofold: 1) we first formalize this semifragile authentication problem by proposing an appropriate payoff: by considering the watermark as an integrity stamp (i.e., its detectability stands as an integrity evidence), a natural game cost is the false-alarm probability which still detects the watermark in a tampered content and 2) this approach is further applied to find the optimal power allocation across parallel additive Gaussian channels of a dithered distortion compensated scalar quantization-based scheme against substitution attacks. The results are specialized to design an efficient discrete cosine transform (DCT) semifragile image watermarking system which allows detecting substitution attacks despite JPEG compression.

Hyperspectral images present some specific characteristics that should be used by an efficient compression system. In compression, wavelets have shown a good adaptability to a wide range of data, while being of reasonable complexity. Some wavelet-based compression algorithms have been successfully used for some hyperspectral space missions. This paper focuses on the optimization of a full wavelet compression system for hyperspectral images. Each step of the compression algorithm is studied and optimized. First, an algorithm to find the optimal 3-D wavelet decomposition in a rate-distortion sense is defined. Then, it is shown that a specific fixed decomposition has almost the same performance, while being more useful in terms of complexity issues. It is shown that this decomposition significantly improves the classical isotropic decomposition. One of the most useful properties of this fixed decomposition is that it allows the use of zero tree algorithms. Various tree structures, creating a relationship between coefficients, are compared. Two efficient compression methods based on zerotree coding (EZW and SPIHT) are adapted on this near-optimal decomposition with the best tree structure found. Performances are compared with the adaptation of JPEG 2000 for hyperspectral images on six different areas presenting different statistical properties.

Assuming imperfect channel estimation, we propose an improved detector for orthogonal frequency-division multiplexing (OFDM) systems over a frequency-selective fading channel. By adopting a Bayesian approach involving the statistics of the channel estimation errors, we formulate an improved maximum-likelihood (ML) detection metric taking into account the characteristics of the channel estimates. As a first step, we propose a modified iterative detector based on a maximum a posteriori receiver formulation which reduces the impact of channel uncertainty on the decoder performance, by an appropriate use of this metric. The results are compared to those obtained by using a detector based on a mismatched ML metric, which uses the channel estimate as if it was the perfect channel. In a second step, we calculate the information rates achieved by both the improved and mismatched ML detectors, in terms of achievable outage rates. These outage rates are compared to those provided by a theoretical (not practical) decoder defined as the best decoder in the presence of channel estimation errors. Numerical results over both uncorrelated Rayleigh fading channels and realistic ultra wideband channels show that the improved detector outperforms the classically-used mismatched approach in terms of bit error rate and achievable outage rates, without any increase in the receiver complexity.

This paper presents a MAP estimation method of CAVLC (Context Adaptive Variable Length Coding) encoded sequences for robust decoding of H.264/AVC video stream. A sequential decoding algorithm jointly exploiting the residual source information (related to VLC codes and source characteristics) along with the MAC layer CRC redundancy in the transmission scheme is proposed. The branch selection metric in the decoding trellis incorporates the usually considered APP weighted by a CRC dependent factor. We also introduce a sub-optimal but hardware realizable version of the proposed algorithm. Significant link budget improvement results have been demonstrated for transmission schemes using BPSK modulated signals sent over AWGN channels.

Zerotree based coders have shown a good ability to be successfully adapted to 3D image coding. This paper focuses on the adaptation of EZW for the compression of hyperspectral images with reduced complexity. The subordinate pass is removed so that the location of significant coefficients does not need to be kept in memory. To compensate the quality loss due to this removal, a signed binary digit representation is used to increase the efficiency of zerotrees. Contextual arithmetic coding with very limited contexts is also used. Finally, we show that this simplified version of 3D-EZW performs almost as well as the original one.

In this paper, we propose an audio watermarking scheme based on the scalar Costa scheme specifically calibrated with a perceptual model allowing to increase the embedding power. For our study, this scheme is designed to be efficient for low bit-rate embedding with sufficient robustness to channel degradations. We present here the main characteristics of our scheme and the way for introducing perceptual models in such a Costa based watermarking scheme without introducing any noticeable artifacts. An evaluation of the robustness of the embedding system is also theoretically discussed by the way of the main channel attack : additive noise from low to high level. We illustrate its relevance in practice using Monte Carlo simulations.

We consider relay-assisted equalization where a half-duplex relay forwards a signal to a receiver in attempt to aid in the task of equalization. All channels are assumed to be frequency selective, and therefore they contribute intersymbol interference (ISI). We first consider the case of a naive amplify-and-forward protocol where the relay forwards a scaled version of its received (and ISI-corrupted) signal. We then consider an equalize-and-forward protocol where the equalizer attempts to perform linear equalization before forwarding its signal. We show that a relay can indeed provide considerable benefit in the task of equalization even when only providing "part-time help", and we demonstrate the performance of the two schemes with simulations.

Consider a base station transmitting information over a downlink wireless communication channel, where the mobiles (the receivers) only dispose of a noisy estimate of the channel parameters, and these estimates are not available at the base station (the transmitter). In this paper, we examine the effects of imperfect channel estimation at the receivers and no channel knowledge at the transmitter on the capacity of the multiuser fading MIMO broadcast channel. We derive the optimal dirty-paper coding (DPC) scheme and its corresponding achievable rates with the assumption of Gaussian inputs. Our results, for uncorrelated Rayleigh fading, are particularly useful for a system designer to assess the amount of training data and the channel characteristics (e.g. SNR, fading process, number of antennas) to achieve target rates. We provide numerical results for a two-users MIMO broadcast channel with maximum-likelihood (ML) channel estimation. These illustrate a practical trade-off between the amount of training and its impact to the multiuser interference cancellation performance. In particular, we observe the surprising result that a broadcast channel with a single transmitter and receiver antenna, and imperfect channel estimation at the receivers, does not need the knowledge of estimates at the transmitter to achieve large rates.

In this paper, we examine the effects of imperfect channel estimation at the receiver and no channel knowledge at the transmitter on the capacity of the fading Costa's channel with channel state information non-causally known at the transmitter. We derive the optimal dirty-paper coding (DPC) scheme and its corresponding achievable rates with the assumption of Gaussian inputs. Our results, for uncorrelated Rayleigh fading, provide intuitive insights on the impact of the channel estimate and the channel characteristics (e.g. SNR, fading process, channel training) on the achievable rates. These are useful in practical scenarios of multiuser wireless communications (e.g. Broadcast Channels) and information embedding applications (e.g. robust watermarking). We also studied optimal training design adapted to each application. We provide numerical results for a single-user fading Costa's channel with maximum-likehood (ML) channel estimation. These illustrate an interesting practical trade-off between the amount of training and its impact to the interference cancellation performance using DPC scheme.

In this work, coding for the relay channel (RC) and the cooperative relay broadcast channel (RBC) controlled by random parameters are studied. In the first channel, the RC, information is transferred from the transmitter to the receiver through a multiplicity of nodes which all "simply" act as relays. In the second channel, the cooperative RBC, each intermediate node also acts as a receiver, i.e., it decodes a "private message". For each of these two channels, we consider the situation when side information (SI) Snon the random parameters is non- causally provided to the transmitter and all the intermediate nodes but not the final receiver, and derive an achievable rate region based on the relays using the decode-and-forward scheme. In the special case when the channels are degraded Gaussian and the side information (SI) is additive i.i.d. Gaussian, we show that 1) the rate regions are tight and provide the corresponding capacity regions and 2) the state Sndoes not affect these capacity regions, even though the final receiver has no knowledge of the state. For the degraded Gaussian RC, the results in this paper can be seen as an extension of those by Kim et al. to the case of more than one relay.

The optimal decoder achieving the outage capacity under imperfect channel estimation is investigated. First, by searching into the family of nearest neighbor decoders, which can be easily implemented on most practical coded modulation systems, we derive a decoding metric that minimizes the average of the transmission error probability over all channel estimation errors. This metric, for arbitrary memoryless channels, achieves the capacity of a composite (more noisy) channel. Next, according to the notion of estimation-induced outage capacity (EIO capacity) introduced in our previous work, we characterize maximal achievable information rates associated to the proposed decoder. The performance of the proposed decoding metric over uncorrelated Rayleigh fading MIMO channels is compared to both the classical mismatched maximum-likelihood (ML) decoder and the theoretical limits given by the EIO capacity (i.e. the best decoder in presence of channel estimation errors). Numerical results show that the derived metric provides significant gains, in terms of achievable information rates and bit error rate (BER), in a bit interleaved coded modulation (BICM) framework, without introducing any additional decoding complexity.

This paper presents a new view of the Bose-Chaudhuri-Hocquengem (BCH) code through the addition of some flexibility to the syndromes distribution in the transmitted sequence. In order to get this flexibility, we derive a necessary condition (NC) on the syndromes distribution for decoding BCH codes, which includes the already known Hartmann-Tzeng proposition. This NC is essentially deduced from the decoding process of BCH code, and is related to the locator polynomial and the requested constraints to guarantee a maximal error-correction capacity. The obtained results have the advantage to be applicable for any considered field (finite or not). Furthermore, we prove that when the correction capacity is equal to 2 or 3, the obtained NC becomes also sufficient. This result is very useful in some practical transmission systems such as orthogonal frequency-division multiplexing systems. Once the pilot tones considered in such systems verify the necessary and sufficient condition, it becomes possible to both reduce the peak-to-average-power rate and correct the impulse noise, present in such multicarrier systems. The usefulness of the presented analysis and the exploitation of the derived condition on the pilot tones distribution is illustrated by simulation results in the case of the Hiperlan2 system

Quantization-based schemes, such as scalar distortion compensated quantization index modulation (DC-QIM), have demonstrated some performance merits in data hiding, which is mainly a transmission problem. However, a number of applications can be stated in terms of the watermark detection problem (also named zero-bit watermarking), and this situation has been seldom addressed in the literature for quantization-based techniques. In this context, we carry out a complete performance analysis of dithered uniform quantizer-based QIM with DC under additive white Gaussian noise. Using large deviation theory, performance is evaluated according to the receiver operating characteristic (ROC) and the total probability of detection error. Under the white Gaussian host signal assumption, scalar DC-QIM is compared with other existing watermarking methods, including quantized projection (QP), spread spectrum (SS) and improved SS (ISS, for which zero-bit watermarking performances with correlation detector are also proposed). Among the compared methods, it is shown that dithered scalar DC-QIM is the more relevant choice to address zero-bit watermarking, due to its host independent performance. A short comparison is also provided with respect to the corresponding transmission problem, thus evaluating the loss in performance due to the detection. We conclude in measuring the performance gain that could be provided by the use of more sophisticated lattice quantizers than the cubic structure

In this paper, a new still image coding scheme is presented. In contrast with standard tandem coding schemes, where the redundancy is introduced after source coding, it is introduced before source coding using real BCH codes. A joint channel model is first presented. The model corresponds to a memoryless mixture of Gaussian and Bernoulli-Gaussian noise. It may represent the source coder, the channel coder, the physical channel, and their corresponding decoder. Decoding algorithms are derived from this channel model and compared to a state-of-art real BCH decoding scheme. A further comparison with two reference tandem coding schemes and the proposed joint coding scheme for the robust transmission of still images has been presented. When the tandem scheme is not accurately tuned, the joint coding scheme outperforms the tandem scheme in all situations. Compared to a tandem scheme well tuned for a given channel situation, the joint coding scheme shows an increased robustness as the channel conditions worsen. The soft performance degradation observed when the channel worsens gives an additional advantage to the joint source-channel coding scheme for fading channels, since a reconstruction with moderate quality may be still possible, even if the channel is in a deep fade.

Quantization-based schemes, such as scalar DC-QIM, have demonstrated performance merits for data-hiding problem, which is mainly a transmission problem. However, a number of applications are stated in terms of watermark detection problem (also named one-bit watermarking), and this situation has been seldom addressed in the literature for quantization-based techniques. In this context, we carry out a complete performance analysis of uniform quantizers-based schemes with distortion compensation (DC) under additive white Gaussian noise. Implementing an exact Neyman-Pearson test and using large deviation theory, performances are evaluated according to receiver operating characteristic (ROC) and probability of error. Optimal DC's regarding to ROC performances are derived. It is pointed out that false-alarm and miss detection capabilities are jointly optimized by the same DC value. Then, performances are compared with raw quantized-schemes (i.e. without DC) and spread-spectrum (SS) watermarking. It is shown that DC-QIM always outperforms QIM and SS for detection task. The gain provided by the DC reaches several orders of magnitude for cases of interest, that is for low watermark-to-noise regimes. A short comparison is also provided with respect to the corresponding transmission problem, thus evaluating the loss in performance due to the detection

Robust video coding has attracted increasing attention during the past few years. This paper proposes a joint source channel coding scheme able to resist transmission errors over mixed Internet-wireless channels. It involves motion-compensated oversampled filterbanks (OFBs). The redundancy introduced by the overcomplete representation in signals at the output of OFBs is employed for error correction of the motion compensated frames. The errors may be due to the wireless part of the channel (random noise), but also to the Internet part (packet losses). The performance of the proposed approach is illustrated for compressed streams transmitted through a packet erasure channel with an averaged packet loss of 6.25% followed by a binary symmetric channel with a crossover probability of 10-2

Hyperspectral sensors have been of a growing interest over the past few decades for Earth observation as well as deep space exploration. However, the amount of data provided by such sensors requires an efficient compression system which is yet to be defined. It is hoped that the particular statistical properties of such images can be used to obtain very efficient compression algorithms. This paper proposes a method to find the most suitable wavelet decomposition for hyperspectral images and introduces the possibility of non isotropic decomposition. The decomposition is made by choosing the decomposition that provides an optimal rate-distortion trade-off. The obtained decomposition exhibits better performances in terms of rate-distortion curves compared to isotropic decomposition for high bitrates as well as for low bitrates

This paper presents a MAP estimator for CABAC encoded data transmitted through a noisy channel. The decoding process has two characteristics (i) it provides an exact result, without approximation (ii) it is compatible with realistic implementations of CABAC in standards like H.264, i.e. taking into account finite precision problems and handling adaptive probabilities and context modeling. Simulation results outline the efficiency of the proposed method when applied in AWGN and UMTS-OFDM frameworks

Direct sequence ultra wide band (DS-UWB) is a strong candidate for high data rate (HDR) wireless personal area networks (WPAN) IEEE 802.15.3a standard for achieving up to 480 Mbps. This paper presents a new system design methodology based on a multicode approach where the spreading sequences have a duration longer than the channel impulse response (CIR) with the goal of achieving higher data rates up to 1.6 Gbps. Multiple codes are allocated to one user and transmitted synchronously in a TDMA access scheme. All codes are modulated by the same QAM constellation, the best combination of number of codes and QAM order for achieving the largest range of coverage is illustrated for 4 target rates: 10, 110, 480 and 1600 Mbps. The selection methodology is based on a theoretical SNIR derivation at the output of a linear minimum mean square error equalizer using classical WPAN channels as specified by IEEE 802.15.3a. Compared to the DS-UWB proposal for IEEE 802.15.3a, the proposed system grants around 3 times range increase

Motivated by increasing interest in energy efficient modulations, we investigate a blind algorithm for biorthogonal signaling. While this modulation has historically been considered only for use in narrowband systems without intersymbol interference (ISI), recent attention has been given to its use in ISI channels. Due to the fact that biorthogonal modulation (BOM) results in a source that is not i.i.d., however, classical adaptive equalization techniques cannot be directly applied to equalization of BOM signals. We review the MMSE and LMS-based equalizers, and then identify some peculiarities that arise in blind equalization of BOM signals when compared to more traditional modulations like BPSK. Next, we present a novel blind algorithm, called TROMBONE, for the adaptive equalization of BOM signals. We discuss the convergence properties of this algorithm, and demonstrate its performance with numerical simulations

In some situations of channel coding with state information (CCSI), the encoder and/or the decoder may not have perfect knowledge of the state information. In these situations, the state information may be viewed as the sum of a dominant (nominal) state information and a relatively weak perturbation. We consider the general case of channel with arbitrary pair of independent and identically distributed (i.i.d), possibly correlated, state information (S1, S2) available at the transmitter and at the receiver, respectively. We first analyze the decrease in capacity, or channel sensitivity to this perturbing noise. Both lower and upper bounds on this channel sensitivity are provided, using Fisher Information. The lower bound turns to be relatively tight, at low Signal-to-Noise-Ratio (SNR), in the Gaussian case, for which we provide a closed form expression of channel capacity degradation. Next, we show that these results can be used so as to increase system immunity to noise, by adapting the encoder to the channel uncertainty. Also, we straightforwardly extend these results to the more practical case where the state information is known only causally at the transmitter. Finally, for illustration purposes, two possible applications in the non-causal and the causal case, respectively, are discussed.

This paper investigates the mathematical framework of the two-dimensional multiresolution analysis adapted to irregularly spaced data. This two-dimensional multiresolution is related on separable multiresolution analysis using non-uniform B-spline functions. For any arbitrary degree of the spline function, we propose an orthonormalization procedure of the scaling and wavelet bases. These orthonormal basis functions satisfy the important features required by a traditional multiresolution analysis such as: (i) the continuity conditions of the scaling and wavelet functions and (ii) the compact support of the scaling and wavelets functions. We show that the orthogonal decomposition is implemented using filter banks where the coefficients depend on the location of the samples on the image grid.

In the integrity checking context of multimedia contents, a malicious user aims at devising a forged content in order to fool a watermarker by making him use as a genuine content. By considering that the watermark acts as an integrity stamp, the false-alarm probability to recover the watermark signature in a forged content is the criterion of interest. We study and solve a game for this criterion between a watermarker and a falsifier which is allowed to perform a substitution attack, i.e. replace the watermarked signal by a non-watermarked content. As for the watermarker, we are concerned with additive spread-spectrum (SS) embedding. Signals are modeled by parallel colored gaussian processes. Due to the intractability of the false-alarm probability, we resort to Chernoff bound as an alternative cost. Our study confirms some common heuristics: the best attacker choice is to substitute the watermarked host signal using a signal which has very close statistics to the original host signal. The best watermarker strategy is to embed the watermark into the weakest frequency power components of the host signal. We finally consider the consequences of these results in terms of frequency embedding domain for an image SS watermarking scheme which has to be robust to compression. This reveals notable differences with informed scalar quantized-based schemes

This paper investigates asymptotically optimal scalar quantizers to address QIM watermark detection with i.i.d. host data and additive noise. False-alarm probability of detection is chosen as the cost to be minimized, keeping the embedding distortion and the miss probability upper-bounded. To avoid the intractability of false-alarm probability, Kullback distance between watermarked and non-watermarked data is adopted instead. The problem is then to seek the quantizer which maximizes the false-alarm error exponent under distortion constraint. Using Lagrange multiplier minimization, a quantizer updating Lloyd-Max-like procedure is used to solve the optimization. For experimental aspects, host data and noise have been set Gaussian. In comparison with uniform or Lloyd-Max quantizers, it turns out that detection performances can be notably enhanced by using proposed application-optimized quantizers. The gain is effective even for small number TV of sample at the detector input. However, this gain becomes more substantial as TV grows. This also emphasises that good quantizers in terms of distortion are not suitable for detection task

In this paper, we present a joint source-channel decoding technique based on exact MAP estimation for data encoded by CABAC (context-based adaptive binary arithmetic coding) in standards like H.264/AVC. Soft decoding is put at work using an improved sequential decoding technique, achieving a trade-off between complexity and efficiency of the proposed algorithms. Error detection is realized by exploiting the binarization scheme and redundancy left in the code string, so that CABAC compression efficiency is preserved and no additional redundancy is compulsory

Due to its high compression efficiency, the H.264 video coder is very sensitive to impairments due to transmission over noisy channels. Most error resilience/concealment techniques provided in the H.264 standard were dealing with packet losses. In wireless environments, the proportion of corrupted packets (and thus considered as lost) may become very high. This paper shows that the H.264 decoder may be seen as a parity check decoder able to detect erroneous packets. Combined with soft estimation techniques, it allows to correct transmission errors and to reduce significantly the number of packets deemed lost. The proposed solution is compatible with the error-resilience features of H.264

In this paper, an improvement of the quantization optimization algorithm for the MPEG Advanced Audio Coder (AAC) is presented. This algorithm, given a bit-rate constraint, minimizes the perceived distortion generated by the signal compression. The distortion can be related to the quantization error level over frequency subbands through an auditory model. Thus, optimizing the quantization requires knowledge of the rate-distortion function for each subband. When this function can be modeled in a simple way, the algorithm can take a one-loop recursive structure. However, in the MPEG AAC, the rate-distortion function is hard to characterize, since AAC makes use of nonlinear quantizers and variable length entropy coders. As a result, the standard algorithm makes use of two nested loops with a local decoder, in order to measure the error level rather than predicting its value. We first describe a partial subband modeling of the rate-distortion function of interest in the MPEG AAC. Then, using a statistical approach, we find a relationship between the error level and the so-called quantization "scale-factor" and propose a new algorithm that is basically similar to a classical one loop "bit allocation" process. Finally, we describe the complete algorithm and show that it is more efficient than the standard one

Many source coding standards (JPEG, H263+, H264), rely heavily on entropy coding with variable-length codes (VLC). However, bitstreams made of VLC-encoded data are particularly sensitive to transmission errors. Recent results tend to use knowledge of the VLC structure in order to perform an efficient decoding of the bitstream. These techniques use a trellis describing the structure of the VLC codebook and assume that some a priori information is available at decoder side. Significant improvements, compared with prefix decoding of bitstreams are achieved. However, the complexity of these techniques may become intractable when realistic VLC codebooks are considered. This paper presents an algorithm for compacting VLC tables. The codewords are grouped into a minimum number of classes. Decoding algorithms may then work on a reduced number of classes, instead of working on the whole set of codewords. A proof of optimality is provided for the VLC table-compaction algorithm. The algorithm is applied to the H263+ VLC codebook and merges the 204 codewords into 25 classes. The resulting compact tables are shown to be exactly equivalent to the initial ones when used with hard decoding algorithms. The properties of the associated soft decoding algorithms using these compact tables are also evaluated.

During the past few years, much effort has been devoted for efficient coding of video. However, this high coding efficiency results in an increased sensitivity to transmission errors. This is the motivation for developing techniques allowing us altogether to transmit video sequences efficiently and to be robust to the transmission errors. This paper proposes a joint source-channel coding scheme using motion-compensated oversampled filterbanks (OFB). Previous work for robust transmission using OFB have concentrated on still images. Here, OFB are combined with motion compensation techniques. The redundancy introduced by the overcomplete representation is employed for error correction of the motion compensated frames. Experiments have shown that performance is acceptable even when the compressed stream is transmitted through a binary symmetric channel with crossover probability of 10/sup -2/.

M-ary bi-orthogonal modulation or M-ary bi-orthogonal keying (MBOK) is a modulation scheme that has recently been considered for use in several consumer wireless standards, including the IEEE 802.11 WLAN standard and the ultra wideband (UWB) 802.15.3a WPAN standard. In this paper, we propose a novel blind adaptive equalizer for such signals. We then discuss the relation of the algorithm to other blind algorithms and demonstrate its local convergence. Finally, we present some numerical examples that demonstrate its features and performance.

The soft decoding of variable-length encoded texture data generated, e.g., by video coders such as H.263+ and sent over a packetized network (Internet or mixed Internet radio-mobile) is considered here. Existing soft decoding techniques usually make use of either the number of bits in a block or the number of symbols in a block. Thus, this side information has to be transmitted. This paper describes an algorithm making use only of information available in the packetized bitstream, and is thus compatible with existing standards. Simulations illustrating the efficiency of the proposed decoding procedure are provided.

This contribution proposes an iterative channel impulse response (CIR) estimation scheme in the context of the pseudo-random-postfix OFDM (PRP-OFDM) modulation. While conventional techniques reduce noise and interference of OFDM data symbols on the CUR estimates by simple mean value calculation, the new proposal uses soft decoder outputs in order to perform iterative OFDM data symbol interference cancellation. In a typical example for 64QAM constellations, the mean-square-error (MSE) of the CIR estimates is improved by approx. 12 dB after three iterations. Based on PRP-OFDM postfixes only, it is thus possible to perform channel estimation for higher order constellations (64QAM and higher) with an initial CIR estimation over a small observation window (for 64QAM typically 30 to 40 OFDM symbols). Any loss in throughput due to additional redundancy for CIR estimation purposes, e.g. pilot tones, learning symbols, etc. is avoided at the cost of an increase in decoding complexity.

Multiple watermarking is concerned with embedding several messages into the same host signal, with different robustness and transparency requirements. This paper proposes two implementable scalar schemes for multiple user "dirty paper coding". The first - straightforward - approach consists of an independent superposition of two scalar dirty paper coding schemes. The second consists in the joint design of a scalar dirty paper coding. This joint approach is based on the ideal dirty paper coding scheme for broadcast channels with noncausal side information known to the transmitter. For this purpose, the "scalar Costa scheme" that has been originally conceived for one user is extended to two users. Performance evaluations, including bit error rates and capacity region curves are provided for both methods, illustrating the improvements brought by a joint design.

Impulse radio (IR) ultra wide band systems (UWB) by nature require an accurate timing. This contribution proposes three digital algorithms for compensating the time jitter resulting from the oscillator driving the receiver sampler. The main idea is to estimate jointly useful data and a fraction of the time jitter by using coloration properties of the oscillator phase noise. A maximum likelihood (ML) estimation based algorithm, a suboptimal but with lower complexity one and corresponding adaptive versions are detailed.

Semi-fragile watermarking can be used to detect illegal and local manipulation of images, while being robust towards legal manipulations such as lossy compression. This paper describes a game-theoretic approach to design a blind semi-fragile watermarking system based on the scalar costa scheme. A natural choice for the game criterion is a false alarm probability, i.e. the probability to still detect the watermark, even if part of the data has been replaced. Optimal embedding strategy is then derived in presence of a legal processing (e.g. lossy compression, through some model). Finally, the results are applied to design an efficient DCT semi-fragile watermarking system which allows to detect local forgery signals in spite of JPEG compression.

Information embedding has recently been recognized as power-limited communication over a "super"-channel with state information non-causally known at the encoder. Based on this equivalence many encoding schemes have been proposed. Quantization index modulation (QIM) family has emerged as an asymptotically information-theoretically optimal embedding function. Optimum transmission rates are achieved when both the watermark-to-noise ratio (WNR) and quantizers dimensionality become sufficiently large. At low WNR however, a trade-off between transmission rate and robustness should be found. This paper is concerned with deriving lattice encoding performances over the resulting modulo lattice additive noise channel (MLAN) for both high and low WNR regimes. To this end, some lattices with good packing and quantizing properties are used and corresponding capacity and bit-error rate (BER) curves are provided for both regular QIM, and distortion-compensated QIM.

This paper studies the evolution of Multiple-Input Multiple-Output (MIMO) system capacity in the context of macro-diversity antenna configurations, with respect to system parameters such as the distance between mobile devices. In this context, we redefine an accurate channel model by proposing a realistic configuration of the RF front-end (RFFE). The impact of this improved model on system capacity is first checked in the micro-diversity context, when comparing our results with the model classically used in the literature. This new model is shown to explain some phenomena observed in actual implementations. Then, in the macro-diversity context, we concentrate on the usefulness of adding antennas, depending on the distance to the transmitter or receiver. The obtained results are strongly impacted by the RFFE model.

A parallel between Reed-Solomon codes in the complex field and multicarrier transmission using orthogonal frequency-division multiplexing (OFDM) is first presented. This shows that when the signal is sent over some channel composed of Gaussian plus impulsive noise, the impulsive noise can be removed by a procedure similar to channel decoding, using information carried by the "syndrome." This result is first derived in a simple situation (oversampled discrete multitone, additive channel), which is merely of theoretical interest. In any case, consecutive zeros, in the output of the OFDM modulator, do not correspond to real subcarriers. Pilot tones are transmitted for synchronization or channel-estimation purposes. These pilot tones are generally scattered among the information ones. Our approach is to use these pilot tones as syndromes, in order to correct impulsive noise. We show that the correction capacity is conditioned by the position of these pilot tones in the transmitted sequence. A protection subsystem based on hypotheses tests is introduced after the decoding operation in order to detect malfunctions of this decoder. The efficiency of this technique is corroborated with simulations in the slightly modified Hiperlan2 context. Other extensions are then provided in order to increase the practical usefulness of the method.

We first provide an analysis of available redundancy remaining in compressed video data. This residual redundancy has two origins: the variable-length code (VLC) syntax and the source constraints. We compute the two in terms of equivalent redundancy bits. Then, a VLC decoding algorithm able to exploit both VLC syntax and source constraints for improving the decoding performance is outlined. Finally, it is shown that by taking this redundancy into account, video decoding over additive white Gaussian noise channels is substantially improved.

Today, multimedia transmission over wireless networks encounters several problems and the result is a poor video quality perceived by the end-user. The paper investigates the feasibility of video streaming transmission over EGPRS mobile networks. Then, a simple system of variable-to-fixed length code is proposed to get a more robust compressed data. Hence, more robust data can be transmitted with less error protected EGPRS modulation and coding schemes which allow higher data rates.

This paper discusses the reduced complexity VLC sequence decoder. The proposed decoding algorithm benefits from all the constraints that can be expressed on the whole sequence. It uses the notion of "survivor sequence" as in the Viterbi algorithm (VA), and applies it to variable length code decoding. The survivor selection takes into account the conventional Viterbi metric, the VLC structure and the source semantics constraints on the sequence. The proposed reduced complexity decoder using the VLC structure projection and intrinsic image properties has been evaluated and compared with the conventional prefix-based VLC decoder, and the existing decoder using only the projection on the VLC structure, and the optimal decoder.

Our work focuses on semi-blind equalization techniques when applied to GMSK-based mobile communication systems. Starting from the works developed in (F. Alberge et al, SPAWC'99), a new semi-blind block algorithm is proposed to provide reliable communication through a typical mobile multi-path channel. The algorithm is shown to have performance similar to that of the Viterbi algorithm in a fixed setting, while being more efficient in time-varying situations.

Transmitting video streams on channels impaired with transmission errors is a very demanding task, mainly when images are predicted from previous ones. In this case, errors on motion vectors can be very harmful. In order to overcome this problem, this paper presents a modified H263+ scheme without motion vector transmission. This is obtained by reestimating these motion vectors at the receiver, based on properties of frame expansions. This procedure is obtained at the cost of an increased bit rate, but shows that robust (and efficient) transmission can indeed be obtained in conjunction with image prediction.

Joint source-channel decoding of variable length codes (VLC) for image and video streaming transmission over unreliable links, such as wireless networks, is a subject of increasing interest. This paper proposes an optimum maximum likelihood (ML) decoder of VLC sequences which exploits additional inherent redundancy in the source information, namely: (i) the correlation between bits, inside a VLC codeword; as well as (ii) the correlation between VLC codewords of a VLC sequence unit (e.g. corresponding to one image block). Performance results for improving video decoding over AWGN channels are then presented and compared to the prefix-based decoder as well as the recently proposed VLC decoders.

The unitarity property of transforms is useful in many applications (source compression, transmission, watermarking, to name a few). In many cases, when a transform is applied on real-valued data, it is very useful to obtain real-valued coefficients (i.e. a reality-preserving transform). In most applications, the decorrelation property of the transform is of importance and it would be very useful to control it under some transform parameter (e.g. in joint source-channel coding). This paper focuses on fractional transforms, as tools for obtaining such properties. We propose a methodology for obtaining them and obtain variants of the discrete fractional cosine (sine) transform which share real-valuedness as well as most of the properties required for a fractional transform matrix. As shown in (I. Venturini et al. IEEE Trans. Signal Proc.), such matrices cannot be symmetric.

This contribution extends the recently proposed pseudo-random-postfix OFDM (PRP-OFDM) transceiver to the multiple transmit multiple receive (MTMR) antennae context. In this paper a new generic MTMR modulator is proposed that grants a low complexity semi-blind identification of all channels between any transmit and receive antenna based on order-one statistics. A new channel tracking strategy is detailed based on a moving average (MA) Doppler modeling that preserves the Jakes channel coefficients correlation properties. Moreover two transceiver architectures are presented with an Alamouti based apace-time (ST) block code: equalization schemes with or without postfix interference cancellation and related modulators. Performance is illustrated by simulations in the context of the 5.2 GHz OFDM WLAN IEEE 802.11a standard.

This paper proposes an iterative joint source-channel decoding of variable length codes (VLC) sequences exploiting the inherent redundancy in the source stream, namely (i) the VLC structure (ii) the sequence structure. An optimum soft-input soft-output (SISO) maximum likelihood (ML) decoder of VLC sequences which is able to exploit both types of redundancies is first proposed, and shown to be optimal. This algorithm is then used in an iterative joint source-channel decoding. Performance results for transmitting video over AWGN channels are presented and compared to the soft-output Viterbi algorithm (SOVA) VLC decoder.

Using electronic watermarks as copyright protection for still images requires robustness against attacks. In this paper we propose a watermarking scheme that is robust to rotation, scaling and translation (RST) distortions. The watermark is embedded in a 1D invariant domain that is a projection of the polar Fourier transform. Roc curves depicting false positives versus detection probability of the watermark are provided, under various geometrical attacks and JPEG coding.

With the increasing use of multimedia technologies, video coding requires higher performance as well as new features. To address this need, the latest ITU-T video coding standard, H.264, has been developed. The motion compensation algorithm is the key of the improvements over H.263+. This paper proposes an H.264-based algorithm which does not require the motion vectors to be transmitted. It extends previous results obtained in the H.263+ context. The main purpose of this work is to check whether our procedure of reestimating the motion vectors at the receiver is compatible with the initial H.264 scheme efficiency. It is shown that this property is obtained at some cost in terms of bit rate for comparable PSNR, but that our new H.264-based scheme has about the same performance level as the plain H.263+ coder. This is essentially a preliminary work, which shows that the most sensitive part or the bit stream can be removed, thus allowing new robust coders to be developed.

Joint source-channel decoding of variable length codes (VLC) for image and video streaming transmission over unreliable links, such as wireless networks, is a subject of increasing interest. Much more residual source redundancy can be exploited when both types of correlation: (i) the correlation between bits inside a VLC codeword as well as; (ii) the correlation between VLC codewords of a VLC sequence unit (e.g. corresponding to one image block), are taken into account. The paper (Hang Nguyen et al, IEEE Int. Conf. on Acoustics, Speech, and Sigpro., 2004) proposes an optimum maximum likelihood (ML) decoder of VLC sequences which can exploit these two kinds of correlation and outperforms the existing VLC decoders. This paper proposes a reduced complexity version of this optimal VLC decoder. Simulation results show that the complexity is reduced by 5 in terms of number of possible candidates, while the performance is only slightly sub-optimal compared with the optimal version of the Hang Nguyen paper.

In this paper the capacity of a specific PPM (pulse position modulation) ultra wide band (UWB) system using a semi-digital demodulator architecture is evaluated. Then the impact of time desynchronization between the local pulse templates and the received signal on the capacity of the link is assessed, the goal being to quantify the jitter effect and delay estimation error on this capacity and issue design recommendations for real system construction.

We propose two related iterative algorithms for computing the capacity of discrete memoryless channels. The celebrated Blahut-Arimoto algorithm is a special case of our framework. The formulation of these algorithms is based on the natural gradient and proximal point methods. We also provide interpretations in terms of notions from information geometry. A theoretical convergence analysis and simulation results demonstrate that our new algorithms have the potential to significantly outperform the Blahut-Arimoto algorithm in terms of convergence speed.

The similarity between Reed Solomon codes in the complex field and multicarrier (OFDM) systems can be useful to correct impulse noise. This similarity is best seen when the input of the IFFT in the OFDM modulator has some consecutive zeros. However, this case is merely of theoretical interest, since consecutive zeros, in the OFDM modulator, do not correspond to real subcarriers. However in many cases, pilot tones are generally scattered among the information subcarriers. Our approach is to use these pilot tones as syndromes in order to correct impulse noise. This paper presents a number of interesting results relating on the distribution of pilot tones in the emitted sequence, where we have a necessary condition under which a maximal correction capacity is guaranteed. This necessary condition is also sufficient in the two simple cases, where the correction capacity is 2 or 3. Efficiency of this technique is illustrated by simulations in the practical context of Hiperlan2.

The paper details a new OFDM modulator based on the use of a pseudo random postfix (PRP-OFDM) and low complexity equalization architectures. The main advantage of this new modulation scheme is the ability to estimate and track channel variations blindly using order one statistics of the received signal. This scheme is thus very well suited in the presence of large Doppler spreads where channel tracking becomes essential. Moreover, the proposal of various equalization structures, derived from the zero padded transmission schemes, allows implementations ranging from low-complexity/medium performance to increased-complexity/high performance.

Joint source-channel decoding of variable length codes for image and video streaming transmission over unreliable links, such as wireless networks, is a subject of increasing interest. We first provide an analysis of available redundancy in compressed video data produced by some commonly used standardized codecs. Then, we show that a significant improvement in decoding performance can be achieved when additional image and video data properties are taken into consideration. Finally, performance results for improving video decoding over wireless channels are presented.

An analogy between oversampled filterbanks and channel codes is considered. Parity-check polynomial matrices and syndromes can be defined. Techniques for detecting and correcting impulse noise are provided. Their performances are illustrated on an example.

We address the problem of image restoration in a multichannel system degraded by unknown blurs and additive noise. To identify the unknown blurs, modelled as FIR channels, we propose a subspace based method which exploits an orthogonality property between signal and noise subspaces. Finally, to demonstrate the performance of the proposed method, we run simulations and compare the results with those of the cross relation method.

In this contribution, we propose a cascade structure to improve the impulse noise cancellation algorithm that we have already presented in F. Abdelkefi, P. Duhamel and F. Alberge [2002]. The new structure of this decoding algorithm is easy to implement and reliable (several parameters can be obtained theoretically rather than tuned experimentally). We also show that this decoding algorithm can be used for eliminating impulse noise and for reducing the peak-to-average power ratio (PAPR) level. The efficiency of this technique is corroborated by simulations in the practical context of Hiperlan2.

In an MPEG Advanced Audio Coder (AAC), for each frequency subband, the quantization parameter called scale factor controls the quantization noise and the bit rate. Tuning this coder would require a characterization of the rate-distortion function per subband, which seems to be available only in a parametric manner. In this paper, we show that an accurate model for the quantization noise, depending on the scale factor value, can lead to a simple expression of the relation between the scale factor and the error power per subband. In other terms, we are now able to obtain an explicit writing of the distortion-rate function. In a second time, we apply these results to the coding algorithm in case of a fixed output bit rate and propose a new coding algorithm. Unlike the one described in the informative annex of the MPEG-2 standard, it makes use of a single loop and happens to be faster and more flexible towards psychoacoustic criteria.

In OFDM systems, pilot tones are emitted for synchronization or channel estimation purposes. These pilot tones are generally scattered among the information ones. Our approach is to use these pilot tones as syndromes, in order to correct impulse noise. We previously proposed (see Abdelkefi, F. et al., ISSPA, 2001) a decoding algorithm in three steps: (1) estimate the amount of impulse noise; (2) seek the location; (3) correct the errors. A protection subsystem was introduced after the decoding operation in order to detect malfunction. Indeed, we added a control step which is able to estimate carefully whether the decoding procedure has worked correctly. We now explain this a posteriori control which is based on the hypothesis test. This approach is new in impulse noise cancellation in multicarrier transmission. The efficiency of this technique is also corroborated with simulations in the practical context of HIPERLAN/2.

This article extends turbo demodulation to the zero-padded OFDM (ZP-OFDM) system by accounting for the noise color and the symbols estimates correlation introduced by equalization. Resorting to realistic simulations, we show that turbo demodulation used with set partitioning labeling can significantly outperform noniterative decoding with Gray labeling. We also show that it increases the performance gap between ZP-OFDM and OFDM with cyclic prefix (CP-OFDM) relative to noniterative decoding, because it amplifies the performance gain due to the guaranteed symbol recovery of ZP-OFDM.

Zero padding (ZP) of multicarrier transmissions has been proposed as an appealing alternative to the traditional cyclic prefix (CP) orthogonal frequency-division multiplexing (OFDM) to ensure symbol recovery regardless of the channel zero locations. In this paper, both systems are studied to delineate their relative merits in wireless systems where channel knowledge is not available at the transmitter. Two novel equalizers are developed for ZP-OFDM to tradeoff performance with implementation complexity. Both CP-OFDM and ZP-OFDM are then compared in terms of transmitter nonlinearities and required power backoff. Next, both systems are tested in terms of channel estimation and tracking capabilities. Simulations tailored to the realistic context of the standard for wireless local area network HIPERLAN/2 illustrate the pertinent tradeoffs.

A parallel between Reed Solomon codes in the complex field and multicarrier transmission using OFDM is first presented. This shows that when the signal is sent over some channel composed of Gaussian plus impulse noise, the impulse noise can be removed by a procedure similar to channel decoding, using information carried by the "syndrome". These results are first derived in a simple situation (oversampled DMT, additive channel), which is merely of theoretical interest. Several extensions are then provided in order to increase the practical usefulness of the method. Simulations combining classical convolutive codes with the above mentioned approach are provided.

A new class of blind equalization cost functions have been proposed (see. Shtrom, V. and Fan, H., IEEE Trans. on SP., vol.46, no.10, p.2674-83, 1998). They have the particularity of being unimodal, and some of them are sectionally convex in the combined system domain (CDS). Both properties seem very attractive for a cost function, though no proof (or specific definition) was given of the sectional convexity. We show, using a particular case of this class, that sectional convexity not only implies fixing a delay, but that it also requires to fix the value of the coefficient associated with such delay. We show that the CMA (constant modulus algorithm) criterion shares the same property, and that this property is difficult to maintain in actual algorithms.

This paper proposes a new joint source and channel coder based on real-valued BCH codes, in which the signal protection (analogous to a signal interpolation) is performed before compression. This allows a better tradeoff in high accuracy compression, since part of the distortion introduced by the compression can be corrected by the "channel decoder". Furthermore, we propose an optimal code allocation procedure, which also allows a good robustness to be obtained when the errors introduced by the channel increase. The resulting rate/distortion curves outperform those obtained by a separate system on the whole range of operation. Although presented in the context of image transmission through a binary symmetric channel, the resulting codes may be employed on a wide range of transmission schemes with significant performance benefits.

In a MPEG-4 analytic audio scene, spatialization filters can be used after an audio coder. When properly tuned, a MPEG advanced audio coder (AAC), generates an inaudible distortion if a sufficient output bit-rate is available, and, more important in the context of our study, as long as the listener has direct access to the unprocessed output of the decoder. We show that using a spatialization filter after the decoder, as described in the MPEG-4 standard, may damage the perceived quality. We propose solutions and present a modified encoder working with a hearing model that includes all audio processing (including spatialization), which allows bit-rate reduction without quality damage.

In this paper, a Bose-Chaudhuri-Hocquenghem (BCH) coder in the field of the real numbers is investigated for simultaneous source coding and impulse noise cancellation of satellite images. Our channel is a binary symmetric channel (BSC). Our approach is to make a carefully designed interpolation of the subband images prior to quantization and transmission. Compared to a classical tandem source and channel coding (TSC) scheme, in which a binary BCH coder would take place after quantization, our approach makes use of BCH coding prior to quantization, thus, allowing joint source and channel decoding. We also examine the issue of robust transmission of wavelet compressed images through the noisy channel: simulations show that we obtain a 3.7 dB improvement in PSNR over the classical entropy coder for a global rate of 3.25 transmitted bits per pixel and small BSC crossover probability.

A block algorithm is presented to solve the joint blind channel identification and blind symbol estimation problem. It is based on a deterministic maximum likelihood (DML) method. A partial prior on the symbols is incorporated into the DML criterion in order to improve the estimation accuracy. We propose a test which permits to circumvent the local minima problem and which is pertinent for a large class of criteria. The structure of the block algorithm is well-suited for deriving recursive and adaptive versions. We prove that, in the noiseless case, the obtained recursive algorithm converges only towards the global minimum. Numerical results show that the prior on the symbols improves the accuracy of the estimators and brings robustness to the lack of channel diversity. At the same time, this method introduces fewer local minima than the use of a full prior.

This paper presents the application of the promising public key watermarking method to the audio domain. Its detection process does not need the original content nor the secret key used in the embedding process. It is the translation, in the watermarking domain, of a public key pair cryptosystem. We start to build the detector with some basic assumptions. This leads to a hypothesis test based on probability likelihood. But real audio signals do not satisfy the assumption of a Gaussian probability density function. Moreover, the use of an advanced human perception model to hide the watermark makes the detection issue a tough problem. Our works result in a new detection process offering a good test's power for a low probability of false alarm.

The widespread application of OFDM for local area mobile wireless broadband systems is strongly motivated by the simple equalization it affords. OFDM's bottleneck of channel-dependent performance was only previously addressed by replacing the cyclic prefix (CP) with trailing zeros (TZ). Such zero-padded OFDM transmissions enable FIR channel-irrespective symbol recovery which results in significant BER gains. The price paid is increased receiver complexity compared to the classical CP-OFDM. To reduce complexity, this paper proposes two novel equalizers for TZ-OFDM. One is based on an overlap and add approach and exhibits exactly the same performance and complexity as CP-OFDM. The second offers BER performance close to the TZ-OFDM MMSE equalizer and incurs a moderate complexity increase. An extension of an existing CP-OFDM pilot-based channel estimation method is also derived for the proposed equalization schemes. Simulations are conducted in a realistic HiperLAN/2 scenario comparing the various OFDM transceivers.

This paper proposes a simple equalizer for a multicarrier block transmission scheme which pads zeros (as opposed to a cyclic prefix) in each transmitted block. In the absence of high-power amplifier induced nonlinear effects, it is shown that the resulting scheme is simply the dual of the classical cyclic prefix OFDM transceiver. Comparison between the two systems that takes into account nonlinear distortions introduced by the clipper is also performed in the practical context of the wireless broadband 5 GHz HiperLAN/2 system. Both the classical pilot-based as well as novel (semi-) blind subspace algorithms are applied for channel estimation. The advantage of subspace methods in lowering the variation of channel tracking is corroborated with simulations which verify practically accurate channel estimates along the bursts.

We investigate a new class of scalable image coders. We target at the same time multiresolution (for spatial scalability) and hierarchical segmentation based decompositions (for object based scalability). Hierarchical segmentation allows to access the description of a scene in terms of regions or objects at several resolution levels, and thus encode and transmit the objects selectively. From a coding viewpoint, it is obviously interesting to couple the multi-level segmentation with a critically decimated decomposition of the image (to avoid redundancy of representation). We propose a new method to evaluate the hierarchical segmentation of an image using its wavelet decomposition. As the segmentation map is embedded by reconstruction inside the decomposition, we do not need to transmit it separately, thus reducing the bit rate. Simulations show that a prototype coder of this type is competitive with conventional image coders, while offering in addition new perspectives for object based coding and manipulations.

Asymmetric schemes belong to second generation of watermarking. Whereas their need and advantage are well understood, many doubts have been raised about their robustness. According to a method presented by Furon and Duhamel (see Proc. of the 3rd Int. Work. on Information Hiding, Dresden, 1999), a very robust symmetric technique is derived into an asymmetric scheme. Tests show that it is as robust as the symmetric version. Yet, asymmetric schemes undergo malicious attacks that confuse the detection process. Tests reveal that the quality loss due to these malicious attacks is too important for the signal to be used after the attack.

Most second order single input multiple output (SIMO) identification algorithms identify the global impulse channel response, convolution of an emission filter and a propagation channel. This paper makes an explicit use of this channel structure in a second order algorithm. We present several structured methods exploiting more or less prior information on the emission filter. Proofs of convergence are provided, and simulations show that some knowledge based algorithms greatly improve over classical blind algorithms, even in the case where the knowledge is partial.

We address the reconstruction problem of a high resolution image from its undersampled measurements across multiple FIR channels with unknown response. Our method consists of two stages: blind multi-input multi-output (MIMO) deconvolution using FIR filters and blind separation of mixed polyphase components. The proposed deconvolution method is based on the mutually referenced equalizers (MRE) algorithm previously developed for blind equalization in digital communications. For sources separation, a method is proposed for separating mixed polyphase components of a bandlimited signal. The existing blind source separation algorithms assume that the source signals are either independent or uncorrelated, which is not the case when the sources are polyphase components of a bandlimited signal. Simulation results on artificial and photographic images are given.

Orthogonal frequency division multiplexing (OFDM) is a promising technique for high data-rate transmission over wireless channels. In general, wireless channels are time-frequency dispersive. The performance of wireless OFDM therefore depends critically on the time-frequency localization of the pulse shaping filter used. It has been pointed out in Haas (1996) that OFDM systems based on offset QAM (OFDM/OQAM) bypass a major disadvantage of OFDM schemes based on ordinary QAM, namely the fact that well-localized pulse shaping filters are prohibited in the case of a critical time-frequency grid where spectral efficiency is maximal. In this paper, we derive general orthogonality conditions for OFDM/OQAM systems and we propose efficient (FFT-based) design procedures for time-frequency well-localized OFDM/OQAM pulse shaping filters with arbitrary length and arbitrary overlapping factors. Finally, we present design examples.

Although zerotree coders are extremely simple in terms of algorithmic complexity and structure, their rate-distortion tradeoff is among the best known in the literature. This paper intends to provide an explanation to this phenomenon, by explicitly showing which part of the algorithm is at its origin. This is somewhat proved by introducing the same mechanism in other type of algorithms, which in turn show a similar rate/distortion performance. More specifically, we compare three coders: The first one is the original zerotree coder, the other ones are two "zerotree-like" coders, where coefficient significance is determined by thresholding (first coder) and by by rate-distortion optimization (second coder). These coders exhibit similar behaviour and performances. Thresholding seems to corresponds to a rate-distortion optimal way of isolating significant data. We show the impact that this procedure has on the distribution of the quantization noise. In particular, thresholding removes the coefficients responsible for the non-uniformity of the quantization noise. The significance map is encoded using classical lossless techniques. All three coders are shown to have (almost) equal rate-distortion characteristics.

Adaptive acoustic echo cancellation in stereophonic teleconferencing is a very demanding application. The characteristics are: a very large number of coefficients, non-stationary input (speech), (slowly) time-varying systems to be identified, plus the specific property that both stereo signals are intrinsically very correlated. Basic versions of stochastic gradient algorithms have difficulties in meeting these requirements. We show that, in a multichannel framework, only a combination of techniques can result in an algorithm which convergence is governed by a quasi-diagonal matrix. Simulations with data recorded in a conference room demonstrate the improvement in convergence of our algorithm compared to the LMS.

Bilinear methods for jointly estimating the channel coefficients and the symbols emitted through these channels are very appealing. However, they can be trapped by local minima. This paper provides: (i) a full characterization of the local minima, (ii) a simple criterion for checking whether the procedure has converged to the global minimum, (iii) a simple algorithm for obtaining this solution, with a proof of convergence.

A new image coding technique based on an Linfinity-norm criterion and exploiting statistical properties of the reconstruction error is investigated. The original image is preprocessed, quantized, encoded, and reconstructed within a given confidence interval. Two important classes of preprocessing, namely linear prediction and iterated filterbanks, are used. The approach is also shown to be compatible with previous techniques. The approach allows a great flexibility in that it can perform lossless coding as well as a controlled lossy one: specifications are typically that only p% the reconstructed pixels are different from the original ones.

This contribution deals with the problem of blind channel identification and equalization based on the (temporally or spatially) oversampled channel output. A novel algorithm is presented which builds on a multistep prediction (MSP) approach and can be viewed as a certain generalization of the initial linear prediction algorithm (LPA) proposed in 1994. Our algorithm improves on related works in that it is theoretically and practically unsensitive to the critical and expected problem of channel length mismatch. Moreover, the MSP scheme improves on the conventional LPA by increasing the robustness of this earlier algorithm. In contrast with the LPA, the proposed prediction scheme exploits the full channel structure, thus providing more statistical efficiency in channel identification. A direct symbol recovery algorithm (requiring no channel estimate) is also straightforwardly drawn from our approach.

The design of filter banks for source coding purposes classically relies on the perfect reconstruction (PR) property. However, several studies have shown that taking the quantization noise into account in the design could yield a noticeable reduction of the mean square reconstruction error. The purpose of this study is to show that perceptual improvement can also be obtained in the particular audio coding context by relaxing the PR constraint. In this context, the mean square error is not relevant any more, and we define a new perceptual distortion criterion, making use of a simplified ear model, the MPE (mean perceptual error). Then, synthesis filters are optimized so as to minimize this MPE. Finally, this MMPE (minimum MPE) filter bank is included in an audio coding scheme. Compared to the corresponding PR filter bank-based scheme by the means of POM (perceptual objective measure), they show an improved audio quality.

This paper is concerned with the blind equalisation of a communication system. We propose to realize the identification of the channel impulse response as well as the noise variance estimation and detection of the emitted sequence of symbols. The signal representation as a hidden Markov model (HMM) has the intrinsic potential for solving such a problem. However, the high performance of such methods is usually obtained at the cost of a high computational complexity and local minima problems. Both issues are addressed in this paper.

This paper presents source coding schemes, based on both parallel and tree-structured filter banks, in which synthesis filters and quantizers are jointly optimized under bit-rate constraint. The design of reconstruction filter banks thus takes into account the amount of subband quantization noise, and minimizes the output mean square error (MSE). All other constraints on the filters, such as perfect reconstruction or maximum selectivity, are relaxed, and noticeable improvements over perfect reconstruction schemes with optimal bit-rate allocation are shown by means of rate-distortion curves, for both synthetic and audio signals.

Image compression using L/sup /spl infin//-norm and confidence interval criteria was shown to have some advantages with respect to flexibility and accuracy. The improvements are due to a local control of the distortion, which allows some local knowledge on the original signal to be exploited. The coding scheme is flexible and can perform a large variety of coding qualities, including lossless compression. This quality is fixed a priori according to the use of the reconstructed image. This paper studies the advantages of the L/sup /spl infin//-norm coding technique applied to medical imaging (mammography), compared to classical L/sup 2/ norm techniques. The coding scheme is recalled and used to achieve different levels of compression: going progressively and accurately from lossless to near-lossless then lossy compression. Results of experiments conducted on several mammograms are presented. When compared to the classical m.s.e based schemes, this approach is shown to give more flexibility and higher quality of the reconstructed mammograms (especially, microcalcifications).

This paper provides a full description of a 2-D discrete cosine transform (DCT) which has the property of requiring only N 1-D DCT of size N for performing the 2-D algorithm. No multiplication is involved elsewhere in the algorithm, thus the multiplicative complexity is halved when compared to classical row/column algorithms. Compared to some other proposed algorithms, our approach has the characteristic of being formally very concise. Once the main result has been obtained, a remaining difficulty is the search for an in-place algorithm. This problem is also solved.

A straightforward generalization of the so-called affine projection algorithm (APA) to the multichannel (MC) case is easily obtained. However, due to the strong correlation between the input signals of the various channels, the resulting algorithm converges very slowly. The article describes the way to overcome this problem and derives an efficient algorithm that makes use of additional orthogonal projections.

Blind channel identification methods based on the oversampled channel output is a problem of theoretical and practical interest. It is first demonstrated that the subspace methods developed in Moulines are not robust to errors in the determination of the model order. An alternative solution is then proposed, based on a linear prediction approach. The effect of overestimating the channel order is investigated by simulations: it is demonstrated that the prediction error method is "robust" to over-determination.

Classical transform-domain algorithms adapt the filter coefficients (in each "frequency bin") by minimizing a criterion depending on a full-band time-domain error. This paper proposes an algorithm which updates each portion of the frequency response of the adaptive filter according to the error in the same subband. For this purpose, a multirate adaptive filter is used where a subband decomposition of the error is performed using critically sampled lossless perfect reconstruction filter banks. This new algorithm is based on the minimization of a weighted criterion by a stochastic gradient algorithm and leads to improvements in the convergence rate compared to both LMS and classical frequency domain algorithms.

This paper proposes an efficient implementation of fast FIR filtering algorithms with useful characteristics for real-time application. They maintain a low processing delay, independent of the filter length. The difficulty is to keep as much as possible of the improvement brought by the reduction of the arithmetic complexity of these fast FIR filtering algorithms without exceeding the Digital Signal Processor (DSP) resources (number of registers, pointers, memory, ...). A particular attention is devoted to the heavy use of pointers which represents a crucial problem. It is solved in this paper by an optimal organisation of data in memory. Improvements of more than 70% in actual timings on an ADSP-2100 compared to the classical algorithm of convolution are obtained, even for very short blocks.

In the context of subband processing of finite-length signals using FIR filter banks, a new technique is derived for achieving exact reconstruction when subband signals are truncated to the same number of samples as the original signal. Using a delayed truncation method in the subbands, it is shown that the missing samples can be recovered exactly by inverting small linear systems. Our approach also applies to time-varying filter banks or wavelet transforms where filters are switched between consecutive input blocks.>

A class of methods for identifying a single input/multiple output finite impulse response system (SIMO-FIR), from the outputs of the system only is presented. These methods rely on a minimal parametric representation of the system solution. They are based on the orthogonality between a 'signal' and a 'noise' subspaces. This is exploited to build quadratic forms whose minimization yields the desired estimates up to a scale factor. It is shown (by numerical simulations) that these methods provide significantly better (in terms of bias and variance) estimates than the method by Tong et al. (1991), while requiring about one half the number of computations. They are thus very attractive for applications, in particular, for narrowband TDMA channel equalization.>

The paper proposes an efficient implementation of fast composite length FIR algorithms on the digital signal processor (DSP) "ADSP-2100", which closely follows the mathematical structure of the algorithm. The difficulty is to keep as much as possible the improvement brought by the reduction of the complexity without exceeding the DSP resources (number of registers, pointers, memory,...). This difficulty was already pointed out in previous papers, in which the algorithm was restricted to a single decomposition due to these constraints. Particular attention was devoted to this problem. The solution is to structure the algorithm in such a way that the organization of data in the memory is optimised. The authors propose an implementation requiring only five pointers whatever the number of iterations. An improvement of more than 50% in terms of actual throughput (number of cycles per point) compared to the implementation of the direct convolution is achieved.>

A new image coding technique using statistical properties of quantization errors and L/sup /spl infin//-norm criterion is investigated. The original image is preprocessed, quantized, encoded and reconstructed within a given confidence interval. We focus on iterated filter banks as a preprocessing technique, and provide a comparison with linear prediction in the case of very good quality (almost lossless) image coding.>

We present a general framework for constructing transforms in the field of the input which have a convolution-like property. The construction is carried out over the reals, but is shown to be valid over more general fields. We show that these basefield transforms can be viewed as "projections" of the discrete Fourier transform (DFT). Furthermore, by imposing an additional condition on the projections, one may obtain self-inverse versions of the basefield transforms. Applying the theory to the real and complex fields, we show that the projection of the complex DFT results in the discrete combinational Fourier transform (DCFT) and that the imposition of the self-inverse condition on the DCFT yields the discrete Hartley transform (DHT). Additionally, we show that the method of projection may be used to derive efficient basefield transform algorithms by projecting standard FFT algorithms from the extension field to the basefield. Using such an approach, we show that many of the existing real Hartley algorithms are projections of well-known FFT algorithms.>

A new block processing algorithm which requires a smaller number of arithmetic operations than the least mean square (LMS) algorithm, while converging faster is described. Theoretical conditions for this convergence to be faster are derived and checked by simulation. Some simulations are provided showing that, at least under some specific conditions, its tracking characteristics are also improved. These performances are obtained by grouping the computations corresponding to a block of successive inputs. Good performances are obtained even with very small blocks.>

A new derivation of recursive adaptive algorithms which leads to a general formulation of the updating equation is provided. After showing that the classical recursive least square (RLS) algorithm is a special case of this approach, it is shown that with some ad hoc approximations this formulation can be applied to the constant modulus (CM) criterion. The algorithm obtained is simulated in the mobile radio context and its performances are compared to those of other classical or more recent algorithms with a constant modulus criterion, showing a significant improvement in convergence rate and residual error.>

Several algorithms are reviewed for computing various types of wavelet transforms: the Mallat algorithm (1989), the 'a trous' algorithm, and their generalizations by Shensa. The goal of this work is to develop guidelines for implementing discrete and continuous wavelet transforms efficiently, and to compare the various algorithms obtained and give an idea of possible gains by providing operation counts. Most wavelet transform algorithms compute sampled coefficients of the continuous wavelet transform using the filter bank structure of the discrete wavelet transform. Although this general method is already efficient, it is shown that noticeable computational savings can be obtained by applying known fast convolution techniques, such as the FFT (fast Fourier transform), in a suitable manner. The modified algorithms are termed 'fast' because of their ability to reduce the computational complexity per computed coefficient from L to log L (within a small constant factor) for large filter lengths L. For short filters, smaller gains are obtained: 'fast running FIR (finite impulse response) filtering' techniques allow one to achieve typically 30% savings in computations.>

Several data compression techniques are in competition for coding high-quality audio signals (sampling rate >or=32 kHz) down to a 64 kb/s bit rate. The authors concentrate on the subband/transform coding scheme using filter banks known as time domain aliasing cancellation (TDAC). This scheme first involves a window on a set of N data, overlapping by N/2 with the previous one, followed by a TDAC transform. First, it is shown that the transform turns out to be a doubly odd DCT, for which a fast algorithm is provided. Then, it is shown how the windowing and the overlap can be merged with the first step of the first algorithm. The total number of operations (multiplications plus additions) required by this algorithm for computing a length-N TDAC plus windowing is exactly N log/sub 2/ N. Implementation considerations are provided.>

A general block-formulation is presented for the LMS (least-mean-square) algorithm for adaptive filtering. This formulation has an exact equivalence with the initial LMS, hence retaining the same convergence properties while allowing a reduction in the arithmetic complexity, even for very small block lengths. Furthermore, tradeoffs between number of operations and convergence rate are obtainable by applying certain approximations to a matrix involved in the algorithm. The usual block LMS (BLMS) hence appears as one of the possible approximations, which explains some of its properties.>

A 2-D DCT (discrete cosine transform) algorithm based on a direct polynomial approach is presented. The resulting algorithm reduces the number of both multiplications and additions compared to previous algorithms. It is shown that, although being mathematically involved, it possesses a clean, butterfly-based structure. Tables comparing the number of operations are provided, as well as flowgraphs.>

An overview of some alternative algorithms for one- and two-dimensional DCTs (discrete cosine transforms) is given. Operation counts are derived for typical examples useful in image processing. It is shown that it is possible to generalize the 2-D schemes to 3-D DCTs as well. The result is that a 3-D DCT can be obtained from a 3-D DFT (discrete Fourier transform) of the same size on reals at the cost of permutations and O(3/2N/sup 3/) multiplications. The scheme involves rotations on eight output points at a time. Improvements through scaling are discussed, and implementation issues (both in hardware and software) are addressed.>

The authors propose a doubling Levinson/Schur algorithm (DLSA) for the superfast solution of real positive definite Toeplitz systems of order n+1, where n=2/sup v/. This algorithm belongs to the class of development the class of Levinson and Schur algorithms based on a doubling (or divide and conquer) strategy, but with the lowest cross point over the initial Levinson algorithm and the lowest arithmetic complexity. The minimum number of multiplications required for this class of algorithm is given. The algorithm was implemented in order to check whether the improvement in arithmetic complexity was hidden by a loss of regularity, resulting in less efficient implementations.>

A relationship between bit-reverse and matrix transpose, in the context of 2-D fast Fourier transforms (FFTs), is explained. This relationship is shown to be useful for hardware and software implementation of 2-D FFTs based on the row-column algorithm: the bit-reverse operations involved in the row and column 1-D transforms of length N can merge with the matrix transpose in-between, to form a single bit-reverse of length N/sup 2/, which is more efficient than considering these operations separately. The relationship is also the basis of a new bit-reverse algorithm.>

Many different fast FIR (finite-impulse response) algorithms are known either based on fast transforms or on fast aperiodic convolution algorithms. Both types of algorithms make use of overlap-add or overlap-save on large data blocks. They lose the multiply-accumulate structure of the FIR filtering process which is best suited for many implementations. To overcome these two problems, a class of fast FIR algorithms, allowing various compromises between arithmetic complexity and regularity of the resulting algorithms, was proposed. It is shown that, by using the Chinese Remainder Theorem on different data lengths, and with different interpolation points, all the main types of algorithms are found, in a unified manner, as special cases of the new one. >

Reasons are suggested why the split-radix algorithm is better than any single-radix algorithm on length-2/sup m/ DFTs (discrete Fourier transforms). The split-radix approach is generalized to length-p/sup m/ DFTs. It is shown that whenever a radix-p/sup 2/ outperforms a radix-p algorithm, a radix-p/p/sup 2/ algorithm will outperform both of them. As an example, a radix-3/9 algorithm is developed for length-3/sup m/ DFTs.>

Small length Discrete Cosine Transforms (DCT's) are used for image data compression. In that case, length 8 or 16 DCT's are needed to be performed at video rate. We propose two new implementation of DCT's which have several interesting features, as far as VLSI implementation is concerned. A first one, using modulo-arithmetic, needs only one multiplication per input point, so that a single multiplier is needed on-chip. A second one, based on a decomposition of the DCT into polynomial products, and evaluation of these polynomial products by distributed arithmetic, results in a very small chip, with a great regularity and testability. Furthermore, the same structure can be used for FFT computation by changing only the ROM-part of the chip. Both new architectures are mainly based on a new formulation of a length-2nDCT as a cyclic convolution, which is explained in the first section of the paper.

Recently, new fast transforms (such as the discrete Hartley transform in particular) have been proposed which are best suited for the computation of cyclic convolution of real sequences. Two approaches using Fourier or Hartley transforms are first compared, showing that the recently proposed FFT algorithms for real data present a lower arithmetic complexity than the corresponding DHT-based approach. Improvements are made to both types of algorithms, leading to different trade offs between arithmetic and structural complexity. We also present a new Hartley Transform algorithm with lower arithmetic complexity than any previously published one.

A new algorithm is presented for the fast computation of the Discrete Fourier Transform. This algorithm belongs to that class of recently proposed 2n-FFT's which present the same arithmetic complexity (the lowest among any previously published one). Moreover, this algorithm has the advantage of being performed "in-place", by repetitive use of a "butterfly"- type structure, without any data reordering inside the algorithm. Furthermore, it can easily be applied to real and real symmetric data with reduced arithmetic complexity by removing all redundancy in the algorithm.

The most promising Number Theoretic Transforms are those with 2 as a root of unity, since they can be performed without multiplications. One of the main problems is then the complexity of the arithmetic modulo M. We present here a generalized form of the NTT allowing the study of the problems of the NTT's and their arithmetic modulo M together. We show that, among one class of NTT's (the moduli being obtained by evaluation of cyclotomic polynomials) there are some relations between the arithmetics involved, that can be used to decompose the "difficult" arithmetics into simpler ones (just like a DFT of length N1N2, (N1,N2) = 1 can be decomposed into several DFT's of length N1and N2). We also point out a possible application to polynomial transforms.

When the specialized hardware is not too severe a constraint, the most promising Number Theoretic Transforms are those with 2 as a root of unity, since they can be performed without multiplication. Unfortunately, for a given wordlength, previously known NTT's with 2 as a root of unity are too short (2^{2}^{t} + 1, 2^{2q}-2^{q}+1) or too long (3.2n+ 1). It is then important to match the modulus of the transforms to the dynamic range of the convolution to be performed. To this end, we propose : - New NTT's (by the fact, new moduli) allowing longer transforms than those modulo2^{2}^{t} + 1or2^{2q}-2^{q}+1, these moduli being obtained through evaluations of generalized cyclotomic polymonials. -An "optimal" algorithm, which further enlarges the possible convolution length.

The multibridge filter (MBF) is a new fully integrated charge transfer device using the concept of passive recirculation of charges. This paper investigates the synthesis problem with the additional constraint of bias conservation in each cell. The general synthesis is solved by use of a conventional minimax algorithm of optimisation.

The following question provided motivation for this paper : - given some frequency specifications to be met by the implemented filter (taking into account the variations of the coefficients), is it better to implement the usual transfer functions with optimum magnitude in the Chebyshev sense, or to implement functions exhibiting a low sensitivity frequency response to the change of the coefficients, perhaps with non-optimum characteristics ? Two methods of synthesis taking into account some measure of the sensitivity, in the case of low-pass (or high-pass) filters are presented. - the first one, in the case of transversal filters, is based on a linear-programming approach. -the second one, in the case of recursive filters, is based on a procedure similar to the one presented in (2), Results demonstrating the advantage of such an approach in the case of CCD recursive filtering are provided.