Examines the problem of segmenting colour images into homogeneous regions for use in content based image retrieval (CBIR) or object recognition in general. Low level features provide intensity, colour and texture characteristics across the entire image. From these feature vectors a measure of local homogeneity is obtained. Through iterative modelling a seed and grow style algorithm is used to locate each segment. The final segment models provide sufficient information for higher level processing or classification. Segmentation and classification results are illustrated from a database of 1000 Corel Photo CD images.

This paper compares the Hough transform and the UpWrite for the detection of lines, circles, and ellipses. Both ideal and noisy images are tested. The UpWrite is found to be more robust for images containing perturbation noise. For ideal images and images with speckle noise, the results are found to depend on the complexity of the object being detected, with more complex objects favoring the UpWrite.

The quantization distortion of vector quantization (VQ) is a key element that affects the performance of a discrete hidden Markov modeling (DHMM) system. Many researchers have realized this problem and tried to use integrated feature or multiple codebook in their systems to offset the disadvantage of the conventional VQ. However the computational complexity of those systems is then increased. Investigations have shown that the speech signal space consists of finite clusters that represent phoneme data sets from male and female speakers and reveal Gaussian distributions. We propose an alternative VQ method in which the phoneme is treated as a cluster in the speech space and a Gaussian model is estimated for each phoneme. A Gaussian mixture model (GMM) is generated by the expectation-maximization (EM) algorithm for the whole speech space and used as a codebook in which each code word is a Gaussian model and represents a certain cluster. An input utterance would be classified as a certain phoneme or a set of phonemes only when the phoneme or phonemes gave highest likelihood. A typical discrete HMM system was used for both phoneme and isolated word recognition. The results show that the phoneme-based Gaussian modeling vector quantization classifies the speech space more effectively and significant improvements in the performance of the DHMM system have been achieved.

The process of detecting lines and curves in an image is an important component of many pattern recognition and computer vision applications. There are well understood approaches to finding curves from a parametrised space of curves. In many practical cases, however, there exist arbitrary shape curves, and different types of curves in the same image. In addition, local perturbation noise can be added to the images. We propose a nonparametric method to overcome the above mentioned problems. The method is based on local modeling of the data. It does not require the specification of a parametric space of curves. It can be used to detect arbitrary curves and thus has a wider applicability than parametric approaches.

The paper describes a speaker-independent isolated word recognition system which uses a well known technique, the combination of vector quantization with hidden Markov modeling. The conventional vector quantization algorithm is substituted by a statistical clustering algorithm, the expectation-maximization algorithm, in this system. Based on the investigation of the data space, the phonemes were manually extracted from the training data and were used to generate the Gaussians in a code book in which each code word is a Gaussian rather than a centroid vector of the data class. Word-based hidden Markov modeling was then performed. Two English isolated digits data bases were investigated and the 12 Mel-spaced filter bank coefficients employed as the input feature. Compared with the conventional discrete HMM, the present system obtained a significant improvement of recognition accuracy.>

The authors present a report on the content and status of a major speech database collection effort. The goal is to collect a useful set of speech material from a very large number of speakers. These speakers are drawn from a wide cross-section of the local community with a variety of ethnic and educational backgrounds. Speech materials include isolated digits and numbers, vowels and voiced phonemes, connected digits, and phonetically balanced sentences. Speech signals are encoded into 16-bit pulse code modulation (PCM) format and stored on Betamax format video tapes. In the seven months since this project started, speech from 100 speakers has been collected. A statistical breakdown of the backgrounds of the speakers is presented.>

The threshold logic unit (T.L.U.) has been proposed as a model for a single neurone; other substantially cognate terms are “perceptron” and “adaline”. Networks of these elements have been advanced as tentative models of some aspects of brain functioning. In particular, hierarchical nets appear to exhibit a sufficient flexibility to make them interesting both as plausible models of learning in the central nervous system and also as general objects of study in connection with pattern recognition and artificial intelligence. In this paper, we discuss the well-known “perceptron convergence theorem” in a fairly general setting, and consider variations appropriate to nets of such units. A certain familiarity with the relevant chapters of Nilsson’s Learning Machines [1] and also with current mathematical formalism is presupposed.