Roger J. Green

Recognition of handwritten cursive Arabic characters

An automatic off-line character recognition system for handwritten cursive Arabic characters is presented. A robust noise-independent algorithm is developed that yields skeletons that reflect the structural relationships of the character components. The character skeleton is converted to a tree structure suitable for recognition. A set of fuzzy constrained character graph models (FCCGMs), which tolerate large variability in writing, is designed. These models are graphs, with fuzzily labeled arcs used as prototypes for the characters. A set of rules is applied in sequence to match a character tree to an FCCGM. Arabic handwritings of four writers were used in the learning and testing stages. The system proved to be powerful in tolerance to variable writing, speed, and recognition rate. >

Fingerprint classification using a hexagonal fast fourier transform

In this paper a new method for the analysis of fingerprint images is presented. A Hexagonal Fourier Transform is applied that will classify fingerprints into whorls, loops and arches. The Hexagonal Fourier Transform allows the utilization of hexagonally sampled data and the extention of output data in a rectangular scheme, which is more convenient for treatment and interpretation. Results for fingerprint classification are given.

Skeletonization of arabic characters using clustering based skeletonization algorithm (CBSA)

Abstract Character skeletonization is an essential step in many character recognition techniques. In this paper, skeletonization of Arabic characters is addressed. While other techniques employ thinning algorithms, in this paper clustering of Arabic characters is used. The use of clustering technique (an expensive step) is justified by the properties of the generated skeleton which has the advantages of other thinning techniques and is robust. The presented technique may be used in the modeling and training stages to reduce the processing time of the recognition system.

Recognition and velocity computation of large moving objects in images

Two one-dimensional time sequences are generated from the projections of the two-dimensional sequence on the x and y axes. Then the two-dimensional fast Fourier transform for the generated time sequences is computed. A peak in the spectrum for the selected spatial frequency is detected. The temporal frequency at which the peak is detected gives an estimate of the velocity of the moving object. Analytical formulations for large moving objects in a time sequence with zero background are presented. An algorithm is given for velocity formulation. >

Bit-error-rate performance of lightwave hybrid AM/OFDM systems with comparison with AM/QAM systems in the presence of clipping impulse noise

Theoretical analysis is presented for the BER performance of hybrid AM/OFDM systems in the presence of clipping impulse and Gaussian noise. In particular, the OFDM BER has been compared with that of conventional QAM. It is found that, due to the dilution effect of the subchannels, the impact of the clipping impulse noise on BER for OFDM systems is virtually zero for reasonably large laser modulation depths. Therefore, OFDM is more suitable to the lightwave transmission of digital video.

Objective methods for evaluation of interferential therapy in the treatment of incontinence

The form of electrotherapy known as interferential therapy is examined with particular application in the management of urinary incontinence. Objective validation of the treatment is described, which allows optimal positioning of the electrodes for a particular patient. It is shown that certain pad positions give rise to higher circulating currents in the pelvic floor, which also corresponds to greater muscle activity as indicated via a pressure probe method. It is shown that objective measurements can be made of muscle activity in sympathy with the difference frequency of two stimulating signals.<<ETX>>

Amplitude density of infrequent clipping impulse noise and bit-error rate impairment in AM-VSB/M-QAM hybrid lightwave systems

Analogue video cavalcade services in cable access television (CATV) networks usually require a very high carrier-to-noise ratio (CNR) in order to provide a good quality picture. In consequence, practical lightwave hybrid amplitude modulation (AM) M-ary quadrature amplitude modulation (M-QAM) systems are found to suffer only infrequent clipping impulse noise, which can be provably modeled as a Poisson arriving pulse train. Based on this knowledge, a new expression for the probability density function (PDF) of the clipping noise at the output of a QAM demodulation matched filter is given which can be numerically evaluated with high accuracy. The bit-error rate (BER) performance prediction is then carried out for M-QAM signals within the hybrid system in the presence of an additive mixture of Gaussian and clipping impulse noise. The agreement between the analytical results and the experimental data is quite good.

Multimedia semantic features and image content description

Content based indexing and retrieval of images and video requires a proper semantic description for image content. The paper discusses the mapping of high level, application specific features to the visual primitives that are accessible through image processing techniques. A major difficulty is that there are currently no established methodologies for describing the contents of an image in terms of semantic features, and we suggest that semiotic approaches could be adapted to the task of image description in limited areas of expertise. After reviewing current trends in the mapping of high level to low level features, we present preliminary results that suggest a possible strategy for mapping semiotic descriptions to image processing primitives.

Colour error reduction in video systems

The problems of achieving color fidelity in video systems are reviewed. Usually a correction matrix is appropriate for one illumination environment, but not necessarily others. A wide range of situations is examined, and, by use of 16 test colors, a matrix can be derived which gives good colorimetry for each. However, as the signal/noise ratio may be degraded, a matrix may be derived which gives a good compromise between noise performance and accurate color rendition. Some color environments are similar, so that ultimately only eight correction matrices are necessary. The noise performance, in terms of the display signal/noise ratio, is determined, and it is shown to be modified by about 1.5 dB for most situations. It is expected that this concept will lead to fully compensating color video cameras. >

The effects of image background on velocity computation of moving objects

Abstract This paper analyses the background effects on identification and velocity computation of moving objects in a sequence of images. The analysis addresses time-varying noise. Computer simulations show the effects of the different forms of the backgrounds on identification and velocity computations. The accuracy of the velocity computations suffers when the noise variance is high. An averaging scheme in the spectral domain is applied to correct and improve the accuracy of these velocity computations.