By Topic

Communications, Speech and Vision, IEE Proceedings I

Issue 5 • Date Oct. 1991

Filter Results

Displaying Results 1 - 18 of 18
  • Removal of subjective redundancy from DCT-coded images

    Page(s): 345 - 350
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (659 KB)  

    The removal of subjective redundancy from video images has recently become an important area of study. A suggested method of removing this redundancy from transform-coded images is through the psychovisual thresholding and quantisation of the image transform coefficients. In this paper, the coefficient thresholding and quantisation levels are based on the combined effects of spatial masking and the varying sensitivity of the human visual system to different spatial frequencies and levels of luminance. By combining the discrete cosine transform (DCT) method of image coding with psychovisual thresholding and quantisation schemes, subdistortion motion video bit-rates as low as 2.5 Mbit/s (non-interlaced 25 frame-per-second video) have been obtained without the need for interframe coding.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Motion detection and estimation of multiple moving objects in an image sequence using the cosine area transform (CAT)

    Page(s): 351 - 356
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (494 KB)  

    A new technique for estimating the velocities of moving objects is presented. This technique is based on applying the cosine area transform (CAT) to the image sequence. The analysis of the CAT domain indicates that the velocities of the moving objects are related to the locations of the peaks in the CAT spectrum. The method of applying the CAT to an image sequence of multiple moving objects is presented. Experimental results are included to demonstrate the applicability of this technique. The technique is simple and gives accurate estimation of the velocities of moving objects. There are no limitations regarding size or velocity of the moving objects.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fast algorithm for VQ codebook design

    Page(s): 357 - 362
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (486 KB)  

    The paper presents a fast codebook training algorithm for vector quantisation. It uses an elimination rule, based on triangular inequality criteria, as well as the partial distortion elimination method, to relieve the computational burden of a conventional codebook training algorithm, including a binary codeword splitting algorithm for the initial codebook and the LBG recursive algorithm. Over 95% savings in both multiplication and addition operations were achieved in the simulation of a VQ codebook training of a 'Lena' image using 16-dimensional vectors.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Comparing coding efficiency of vector chain coding and run-length coding for line drawings

    Page(s): 363 - 370
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (649 KB)  

    More efficient data compression can be achieved in encoding line drawings by vector chain coding (VCC) than by the traditional run-length coding (RLC), provided the total length of the lines within a line drawing is not excessive. More bandwidth or time can thus be saved in transmitting such pictures by using VCC. Although this has so far been established only intuitively, quantitative analysis and comparison of the coding efficiency of these two codes for line drawings are performed in this paper. The coding efficiency is measured in terms of both per-length coding rate and data compression ratio, which are determined for a class of handwritten line drawings characterised by a proper statistical model. In particular, the critical point of line drawing complexity is derived at which VCC becomes less efficient than RLC. Experimental observations are also presented to verify the theoretical results.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Border marriage: matching of contours of serial sections

    Page(s): 371 - 376
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (523 KB)  

    Surface triangulation methods are often used to reconstruct biomedical objects from serial tomographs. The contours extracted from such images must be matched prior to surface tiling. Previously, researchers used relatively simple objects with trivial matching conditions, or used a priori knowledge, or matched the contours manually. This paper describes a new method for matching contours based on topological considerations. This method can automatically tile complex objects; particular attention is placed on objects with hollow interiors. The method also detects complex hole bifurcations and generates appropriate saddle points. Applied to biomedical modelling, this method can match complex contours with a high degree of accuracy.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Karhunen-Loeve method for data compression and speech synthesis

    Page(s): 377 - 380
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (336 KB)  

    The use of the Karhunen-Loeve (KL) method in speech data compression and synthesis using the Fourier-Bessel (FB) expansion coefficients of speech signals is described. Bessel functions seem to make a natural basis for speech signal decomposition. Sinusoidal functions are the eigenfunctions of vibrating strings. Bessel functions are the eigenfunctions of vibrating pipes. The vocal tract resembles an excited pipe rather than a vibrating string. Good quality intelligible speech signals can be reconstructed using only a small portion of the FB expansion coefficient. Further data compression is possible through KL transformation of the speech signal FB expansion coefficient for efficient speech coding and synthesis. The transformation is implemented by first forming a covariance matrix of the FB coefficients. Eigenvalues and eigenvectors of the covariance matrix are computed and ranked according to the eigenvalue magnitude. Speech signals are then reconstructed using only the feature corresponding to the larger magnitude eigenvalues of the covariance matrix.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Combination of block-coded modulation and trellis-coded modulation

    Page(s): 381 - 386
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (399 KB)  

    It is known that trellis-coded modulation (TCM) can give a significant coding gain without compromising bandwidth efficiency, by the concepts of signal set partition and trellis coding scheme. It is also known that block-coded modulation (BCM) can be used to approach closely the maximum energy saving that can be obtained by using the densest lattice over the rectangular lattice (QAM). In this paper a systematic scheme is proposed to combine TCM and BCM, such that the constructed coded modulation is similar to part of Wei's multidimensional TCM (1987). The main idea is that the signal points that are used in multidimensional TCM should be partitioned according to the dense lattices formed by proper BCM. Hence a higher coding gain is obtained which is a combination of both the distance gain of TCM and the density gain of BCM. In the 2n-dimensional case, for n>or=4, the net asymptotic coding gain of this scheme can theoretically reach (6-3/n) dB.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Optimal design parameters in capacity assignment for a broadband network

    Page(s): 387 - 393
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (447 KB)  

    A capacity assignment strategy for a broadband network is proposed and studied in this paper. This broadband network has a large channel capacity in its elements. It usually has two kinds of customers: one is queueable narrowband (NB), and the other is blockable wideband (WB) which needs real-time delivery. Besides the characteristics of these two types of traffic, since the improvement in the blocking probability of WB is more significant than that in the mean waiting time of NB, the capacity assignment strategy favours the WB traffic by providing bit-rate compression, reserving some channels for WB customers with non pre-emption by NB customers, and permitting NB customers to use free WB channels only when the NB queue exceeds a certain length. An analytic model is developed and the solutions are presented. The results show that, via proper choice of design parameters, the proposed strategy can have superior performance.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Solution for packet switching of broadband ISDN

    Page(s): 394 - 400
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (664 KB)  

    A possible solution for the broadband ISDN is proposed. A new packet switch, denoted by DPPS (double phase packet switch), is found to be capable of performing well with low cost. DPPS consists of several baseline networks in parallel. DPPS is modular and can easily be expanded because of the recursive structure of baseline networks. Along with the simple and distributed control algorithm, the DPPS can be implemented in VLSI and is consequently high speed and high capacity. Analysis results show that DPPS is able to achieve good throughput-delay performance even under heavy traffic loads.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Synchronisation of random bit sequences: application to hitless switching

    Page(s): 401 - 406
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (431 KB)  

    The paper deals with the synchronisation problem of random bit sequences in hitless, or errorless, switching in frequency diversity digital radio systems. After a short discussion of technical aspects of the switching system, the problem of synchronisation is defined. The sources of incorrect operation, i.e. of not finding synchronism and of finding it in a false position, are given. It is shown that in both error free and highly erroneous situations bit sequences are synchronisable, in the statistical sense, if the delay difference is 'low', as defined in the paper. In the noisy case an error probability lower than 0.5 is the only restriction. Statistics on the synchronisation delay and the necessary period of observation are given.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Decomposition of output queuing switches

    Page(s): 407 - 416
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (789 KB)  

    The paper introduces a simple technique to reduce the complexity of the output queuing switch with a speedup without either losing the self-routing characteristic or allowing performance degradation under nonuniform traffic patterns. Any of the output queuing switches with a speedup can be used as a basic switching element. The switching elements are then interconnected with the distributors in multiple stages. By introducing the distributors between stages of switching elements, the packets can be evenly distributed to all the input ports of the switching elements in the following stages. Although the switch becomes blocking, it retains the self-routing property and achieves the maximum throughput of 100% with only a small additional delay.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Distributed double-phase switch

    Page(s): 417 - 425
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (710 KB)  

    A packet switch architecture is proposed for the broadband ISDN. It is in three stages. The first stage consists of simple distributors which transfer the incoming packets to the later stages regularly and in balance. The next two stages carry out the routing; they are composed of small knockout switches. The authors model the switching system as an M/D/m queuing model and show that, when the load is under 0.84, the following features are provided in a service-independent communication: the probability of losing a packet is less than 10 -6; and the average delay of packets is less than three packet time. The switching system can offer higher load in a service dependent communication while maintaining the same transmission quality. They evaluate the cost and throughput of the design in terms of today's VLSI technologies. For a typical local switch centre whose size is less than 16384*16384, the cost is 2 chips per port and a line rate of 150 Mbit/s is achievable.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Analysis and simulation of a hybrid modulation scheme for embedding data in amplitude-modulated signals

    Page(s): 426 - 436
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (881 KB)  

    Zero synchronous modulation (ZSM) is a novel hybrid modulation scheme which employs angle modulation to transmit a data signal in addition to analogue envelope modulation. The bandwidth of a ZSM signal is restricted to that of conventional DSB-AM by synchronising data with the occurrence of complex zeros of the analogue input. The theoretical basis of ZSM is discussed and a new algorithm is given for locating complex zeros in real time. Analysis of ZSM is extended to implementation errors at the transmitter due to inaccurate zero location and finite truncation of angle modulating functions. The special case of sinusoidal envelope modulation is discussed. Simulation results illustrate the performance of ZSM under practical operating conditions with sinusoidal, random and voiced speech analogue input. Power spectra are computed which indicate that average data rates of about half the bandwidth of the analogue envelope modulating signal can be sustained without a significant increase in bandwidth over conventional DSB-AM.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Sounding techniques for wideband mobile radio channels: a review

    Page(s): 437 - 446
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (806 KB)  

    A major prerequisite to the specification and design of any wideband digital mobile radio system is a thorough knowledge of the propagation characteristics of the mobile radio channel. The various methods of channel sounding are reviewed, with both narrowband and wideband techniques being discussed. The advantages and limitations of each method are highlighted and the swept time-delay cross-correlator (STDCC) technique is presented as the optimum choice. The main features of an experimental STDCC sounder are identified, factors that are considered essential to good system performance are discussed, and the interrelationships between the sounder parameters are emphasised. It is shown that extension of the power-delay profile prior to calculation of the frequency correlation function leads to the achievement of substantially improved frequency resolution. This is particularly relevant when the average power delay profile contains significant echoes at large excess time delays.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Measurement and modelling of wideband mobile radio channels at 900 MHz

    Page(s): 447 - 457
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (882 KB)  

    The complex bandpass impulse response of urban mobile radio channels is measured at 900 MHz using a swept time-delay cross-correlator channel sounder. The propagation data are analysed with particular relevance to the development of a representative channel simulation and production of parameters useful to systems engineers. In both cases, a two-stage characterisation of the data is used. A moving average normalisation method is applied to the processing of the simulation data. The wideband propagation path is shown to be well modelled by a tapped-delay line, with each tap comprising an uncorrelated Rayleigh modulator, a partially correlated log-normal modulator and a weighting factor. These results lend support to the Gaussian wide-sense stationary uncorrelated scattering model. The simplicity of using this model is highlighted by way of an example. Gross non-stationarity in the channel impulse responses identifies the need for formulating cumulative distributions of the small-scale systems parameters. It is shown how these results may be used in assessing the performance of existing and future mobile radio systems.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Performance analysis of mobile packet radio networks in real channels with inhibit-sense multiple access

    Page(s): 458 - 464
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (449 KB)  

    Coherent cumulation of interference signals is modelled in an environment of Rayleigh fading and log-normal shadowing under the influence of near-far effect. The interference model is used to investigate the performance of mobile packet radio networks using the inhibit-sense multiple access protocol. Numerical results are presented, indicating the effect of individual and combined propagation impairments on channel capacity. It is shown that the combined effect of propagation impairments of the real channel enhances the channel capacity.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Bit error probabilities of CPM signals in frequency-selective fading channels

    Page(s): 465 - 472
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (604 KB)  

    The performance of continuous phase modulation (CPM) signals with an MSK-type receiver in the frequency-selective fading environment is investigated. The channel is characterised by its power delay profiles. The modulation index of CPM signals is restricted to 1/2 for coherent detection with an MSK-type receiver. The full system performance is analysed analytically and the numerical results for the bit error probabilities of MSK, GMSK, 3RC and TFM are provided. The effects of changing the transmission rates and the power delay profiles are also discussed. It is that GMSK shows lower bit error probability than MSK in a frequency-selective fading channel.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Differential detection of correlative encoded continuous phase modulation schemes using decision feedback

    Page(s): 473 - 480
    Save to Project icon | Click to expandQuick Abstract | PDF file iconPDF (537 KB)  

    The authors generalise a differential detection technique, introduced elsewhere for a Gaussian minimum shift keying (GMSK) scheme, to include any correlative encoded continuous phase modulated (CECPM) signalling format. The proposed symbol-by-symbol receivers employ decision feedback to partially remove the effects of the destructive intersymbol interference (ISI) which corrupts the differentially detected CECPM signal. After achieving wider eye opening with the decision feedback, the outputs of the two or more differential detectors are jointly utilised to further improve the performance. As a typical example of the CECPM family of signals and because of its excellent spectral characteristics, the new receiver configurations employing up to 3-bit differential detectors have been applied to a tamed frequency modulation (TFM) signal. Bit error rate (BER) performance evaluation results have indicated that performance improvements of more than 5 dB (at BER=10 -4) over a conventionally differential detected TFM scheme are possible.<> View full abstract»

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