By Topic

Vision, Image and Signal Processing, IEE Proceedings -

Issue 3 • Date Jun 1994

Filter Results

Displaying Results 1 - 12 of 12
  • Speaker-independent isolated word recognition using multiple hidden Markov models

    Page(s): 197 - 202
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (440 KB)  

    A multi-HMM speaker-independent isolated word recognition system is described. In this system, three vector quantisation methods, the LBG algorithm, the EM algorithm, and a new MGC algorithm, are used for the classification of the speech space. These quantisations of the speech space are then used to produce three HMMs for each word in the vocabulary. In the recognition step, the Viterbi algorithm is used in the three subrecognisers. The log probabilities of the observation sequences matching-the models are multiplied by the weights determined by the recognition accuracies of individual subrecognisers and summed to give the log probability that the utterance is of a particular word in the vocabulary. This multi-HMM system results in a reduction of about 50% in the error rate in comparison with the single model system View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Error activity and error entropy as a measure of psychoacoustic significance in the perceptual domain

    Page(s): 203 - 208
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (436 KB)  

    Several models have been described in the literature which seek to represent audio stimuli in the perceptual domain to best predict the audibility of errors and distortions. By modelling the principal nonlinear processes of human hearing it is possible to calculate a perceptual domain error surface that represents the audible difference between distorted and original audio signals. A further stage of analysis is required to maximise the usefulness of the auditory model output. The audible error surface must be interpreted to produce an estimate of the overall subjective judgement which would result from the particular distortion. Ideally, the interpretation of the error surface should be broadly analogous to human perceptual mechanisms, and equally, it would be desirable to avoid the complex and cumbersome statistical mapping and clustering techniques proposed by some authors. A technique employed in adaptive transform coding of images, namely cell entropy, offered several desired properties. The paper reports the extension and application of such a technique to the interpretation of perceptual-domain error surfaces produced by an auditory model. Speech data were subjected to an example, algorithmically generated, nonlinear distortion and then processed by the auditory model. The usefulness of the error-activity and error-entropy quantities are illustrated, without optimisation, by comparison of model predictions and experimentally determined opinion scores View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multiplier-less FIR filter design using a genetic algorithm

    Page(s): 175 - 180
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (396 KB)  

    A new synthesis technique is described for multiplier-less FIR digital filters consisting of a cascade of primitive linear phase sections. For medium-order filters the search space for an optimal cascade is typically of the order of 1030 and this can be examined in a computation efficient way using the parallel-search capability of a genetic algorithm (GA). A particular form of GA based upon multilevel or structured chromosomes has been developed for the primitive cascade problem. Initial results suggest that, for the cost of increased filter delay, a typical 2:1 advantage can be achieved in both VLSI chip area and clock rate compared to filters designed using the usual equiripple method View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Enhanced channel coding using source criteria in speech coders

    Page(s): 191 - 196
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (352 KB)  

    Spectral efficiency in digital voice communications in personal communications networks is primarily provided for by advanced speech coding techniques. A threat to the quality and low bit rate requirements of these speech coders is imposed by the transmission channel. Channel coding is thus considered mandatory, but its performance is limited to specific channel conditions and system constraints like bandwidth and transmission power. The authors describe a technique for exploiting source redundancy in speech coders for further improving the performance of the channel coding scheme without the obligatory increased redundancy View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multiresolution morphological fusion of MR and CT images of the human brain

    Page(s): 137 - 142
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (484 KB)  

    A hierarchical image fusion scheme is presented which preserves the details of the input images regardless of their scale. The technique is demonstrated by fusing images of the human brain derived from magnetic resonance (MR) and computed tomography (CT) scanners. Results are given to show that fused images preserve a more complete representation of anatomical and pathological structures, providing information that cannot be obtained by processing the images at a single scale View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • FIR filter order reduction: balanced model truncation and Hankel-norm optimal approximation

    Page(s): 168 - 174
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (392 KB)  

    Two different algorithms for approximating FIR by IIR filters are treated: truncation of the balanced model and the Hankel-norm optimal approximation. Both are assessed for approximation fidelity, as well as for intrinsic computational efficiency. Examples show surprisingly good relative performance of the balanced model truncation, suggesting that frequently this method will be operationally preferable View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Recursive single-most-likely-replacement channel equaliser

    Page(s): 185 - 190
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (420 KB)  

    The paper proposes a recursive single-most-likely-replacement (SMLR) equaliser, that is a fixed-lag block signal processing algorithm indexed by the block size L and the number of decisions N⩽L at each recursion, for channels in the presence of intersymbol interference of finite or infinite length and additive white Gaussian noise. Both computational load and storage required by the proposed recursive SMLR equaliser are linearly proportional to the block size. Two simulation examples illustrate the performance of the proposed recursive SMLR equaliser View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Direct approach to design of PCAS filters with combined gain and phase specification

    Page(s): 161 - 167
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (392 KB)  

    New approaches have been reported for the design of recursive digital filters with prescribed gain and delay using classical optimisation, simulated annealing and genetic algorithms. The filter structure used consisted of two allpass subfilters in parallel (PCAS). PCAS filters are used because of their low complexity and roundoff noise as well as their ability to realise nonminimum-phase transfer functions. An approach is described in which a set of linear equations are formed for each subfilter using a result due to Gregorian and Temes (1978). By imposing certain constraints on these equations, which relate to the filtering characteristics required, they may be solved yielding the coefficients of the two transfer functions. Several examples are given for the case of approximately linear phase. The L2 norm is used as a metric to enable comparison with optimisation techniques. General comments are made on the relationship between the various parameters such as ap, as, filter order, phase slope, etc. The design of PCAS filters with arbitrary phase is discussed View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Statistical analysis of state-space approach in harmonic retrieval

    Page(s): 181 - 184
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (240 KB)  

    The author presents a statistical analysis of the performance of the state-variable balancing for estimating the parameters of exponential signals in the presence of additive noise. The case of frequency estimation for a single damped sinusoid is carried out in detail because the analysis can be presented more clearly; the formulas are simpler and provide insight; and the results are applicable in radar, sonar, communication, and data modelling. Analytical expressions for the variances of the frequency estimate at high signal-to-noise ratios is derived. Both analytical and experimental results are presented to illustrate the performance of the state-variable balancing method View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Low bitrate video coding with depth compensation

    Page(s): 149 - 153
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (696 KB)  

    A depth-compensated low bitrate hybrid video coder for videophone applications is described. Depth and associated position information is extracted from the video frames using an edge-based stereo algorithm. A global motion vector and scale factor are extracted from the depth map and used as parameters for a model of the moving object. These global parameters are used to make the motion-compensated prediction more effective, by compensating for the change in size of the object. Simulations show that the compensated coder results in coded sequences with SNRs up to 1 dB better than those coded with conventional hybrid coders, at a coding rate of 64 kbit/s View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fast closest codeword search algorithms for vector quantisation

    Page(s): 143 - 148
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (424 KB)  

    One of the most serious problems for vector quantisation is the high computational complexity of searching for the closest codeword in the codebook design and encoding phases. The authors present a fast algorithm to search for the closest codeword. The proposed algorithm uses two significant features of a vector, mean value and variance, to reject many unlikely codewords and saves a great deal of computation time. Since the proposed algorithm rejects those codewords that are impossible to be the closest codeword, this algorithm introduces no extra distortion than conventional full search method. The results obtained confirm the effectiveness of the proposed algorithm View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Novel approach to the design of I/Q demodulation filters

    Page(s): 154 - 160
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (440 KB)  

    A novel filter design approach to digital I/Q demodulation is proposed. Two possible realisations are presented using this approach. The first one is based on the highpass filter method which is suitable for B⩽f0 and while the other realisation is based on the lowpass filter method suitable for B⩽f0, where B and f 0 are the IF signal bandwidth and the IF frequency, respectively. Both new realisations maintain the advantages of an earlier lowpass approach such as zero DC offset, matched channel frequency responses, and good performance over a wide bandwidth. At the same time, the new highpass filter realisation method possesses higher computational efficiency than other wideband approaches reported in the literature View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.