Audio-visual speech recognition employing both acoustic and visual information is a novel extension of it significantly improves the accuracy in noisy environments. Although various audio-visual speech-recognition systems have been developed, rigorous detailed comparison potential geometric features from speakers' faces essential. Thus, this paper are compared analyzed rigorously for their importance recognition. Experimental results show that among analyzed, lip vertical aperture most...

Audio-visual speech recognition is a novel extension of acoustic and has received lot attention in the last few decades. The main motivation behind bimodal characteristics perception production systems human beings. effect modeling parameters hidden Markov models (HMM) on accuracy recognizer analyzed, comparative analysis different HMMs that can be used presented, finally model, which been experimentally verified to perform better than others proposed. Also, geometric visual features are...

In this letter, we propose an iterative belief propagation (BP) channel detector (equalizer) over single-input single- output (SISO) and multiple-input multiple-output (MIMO) frequency selective fading channels as alternative to the typically used maximum <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a-posteriori</i> (MAP) or likelihood (ML) detectors. The proposed has a parallel structure, resulting in fast hardware implementations. Moreover,...

We discuss a type of hidden Markov model (HMM) based on fuzzy sets and integral theory which generalizes the classical stochastic HMM. The Choquet is used as relaxes one two independence assumptions that we had with apply this new to speech recognition compare performance In research, main innovation generalized HMM applied for first time recognition. Due fuzziness model, an interesting gain can be observed in terms lower computation time.

In this paper, we propose an iterative belief propagation (BP) based equalizer for frequency selective fading channels as alternative to the typically used maximum a-posteriori (MAP), Viterbi, or soft output Viterbi algorithm (SOVA) equalizers. The proposed has a parallel structure unlike trellis equalizers and it is less complicated than MAP equalizer. Via simulations show that, over channels, BP achieves optimal performance, even in presence of length 4 cycles. We also discuss joint...

We introduce iterative noise whitening for belief-propagation-based channel detectors over intersymbol interference channels with correlated additive noise. In particular, we consider longitudinal magnetic recording systems. The proposed "noise predictive belief propagation" (NPBP) detector iteratively whitens the samples by modifying edge probability computation. NPBP based on finite impulse response and infinite prediction filters are developed. addition, a novel filter optimization method...

In this paper, we propose an iterative belief propagation (BP) detector over multiple-input multiple-output (MIMO) frequency selective fading channels as alternative to the typically used maximum a-posteriori (MAP) or Viterbi detectors. Unlike trellis based detectors having serial structure, proposed has advantage of a parallel resulting in more efficient and fast hardware implementations. Moreover, BP is less complicated than MAP it short decoding delay which essential for reduced memory...

The use of soft computing for virtual object implementation is a novel technique that has received lot attention in the last few years. This field very useful especially applications related to computer games and environments. One application areas real time lip tracking information lips animated characters. increased power computers made it possible do tracking. can be used implement control lip. On other hand, represent parameters enables us have more robust flexible system which...

In this paper, a tensor product-based classification code (CC) is proposed to generate coarse soft information based on the hard flash reads for use in information-utilizing error correction codes (ECCs), particularly low-density parity check (LDPC) codes. Analysis and simulation results have shown that with small ECC overhead, CC helps LDPC decoder achieve better performance compared hard-input case. Furthermore, has very simple encoding decoding algorithms; therefore, it can be easily...

Bimodal speech recognition is a robust technique for automated analysis, and has received lot of attention in the last few decades. In this paper, we analyze effect HMM models on performance bimodal recognizer, present comparative analysis different that can be used recognition, finally propose novel model, which been experimentally verified to perform better than others. One unique characteristic our model fusion strategy acoustic visual features, takes into account sampling rates these two...

High bit error rates of next-generation flash devices necessitate the use more powerful correction codes (ECCs), such as low-density parity-check (LDPC) codes, instead legacy Bose-Chaudhuri-Hocquenghem (BCH) codes. Unlike algebraic random nature LDPC well their ability to soft information requires Monte Carlo (MC) simulations evaluate code performance. Given a large volume NAND data, this can pose resource challenges both in terms simulation platforms and time needed for simulations. In...

Although turbo- and low-density parity check (LDPC) codes provide very low bit-error rates, when these inner are used in conjunction with an outer error correction code (ECC), the overall recording system performance is based on sector failure rates. Therefore, it crucial to identify blocks having more errors than ECC capability. In this paper, we introduce several "burst identification" techniques for turbo LDPC codes. We discuss effectiveness drawbacks of methods over additive white...

Inter-symbol interference (ISI) channels are encountered in many applications, including magnetic recording and wireless communication systems. In general maximum a-posteriori (MAP) or likelihood (ML) algorithms used for channel detection. Recently, iterative belief propagation (BP) algorithm is proposed detection as well. When factor graph of the ISI length 4 cycle free, BP gives a similar performance with MAP detector, while it has advantage parallel structure leading to advantages read...

In this paper, we introduce iterative pattern-dependent noise prediction for belief propagation (BP) channel detectors over intersymbol interference channels with correlated noise. The new scheme, called predictive propagation, makes use of factor graphs, and it iteratively whitens the samples by modifying edge probability computation BP algorithm. To illustrate benefits proposed specifically, consider longitudinal recording corrupted data-dependent media Simulation results convergence...

In this brief, we propose a simple bit error rate (BER) estimator utilizing the parity check constraints of low-density parity-check code-based correction coding (ECC) scheme. The proposed can be efficiently implemented on flash memory die with very small area cost. We articulate applications in which controllers take advantage having BER estimate before ECC decoding scheme to significantly reduce programming errors and improve system performance terms latency throughput.

We introduce iterative noise whitening for belief propagation (BP) based channel detectors over intersymbol interference (ISI) channels with correlated noise. Called predictive (NPBP), the new scheme iteratively whitens samples by modifying edge probability computation of BP algorithm. NPBP on finite impulse response (FIR) and infinite (IIR) prediction filters are introduced. In addition, we propose a novel filter optimization method leading to better performance. Simulation results both...

We study the impacts of channel reliability information (specifically packet error rate (PER)) on topology control (TC). Realizing its importance, we propose use PER along with distance that is used by existing TC mechanisms. considered both centralized and distributed versions when combining distance. Simulation results demonstrate significantly improves energy-efficiency, particularly re-transmission threshold high.