This work is supported by EPSRC research grants, HMGCC, TUBITAK researcher exchange grant, the IST Programme of the European Community, under the PASCAL2 Network of Excellence, IST-2007-216886.

This paper aims to improve the performance of an HMM-based offline Thai handwriting recognition system through discriminative training and the use of fine-tuned feature extraction methods. The discriminative training is implemented by maximizing the mutual information between the data and their classes. The feature extraction is based on our proposed block-based PCA and composite images, shown to be better at discriminating Thai confusable characters. We demonstrate significant improvements in recognition accuracies compared to the classifiers that are not discriminatively optimized.

Fast and accurate analysis of fluorescence in-situ hybridization images for signal counting will depend mainly upon two components: a classifier to discriminate between artifacts and valid signals of several fluorophores (colors), and well discriminating features to represent the signals. Our previous work (2001) has focused on the first component. To investigate the second component, we evaluate candidate feature sets by illustrating the probability density functions and scatter plots for the features. The analysis provides first insight into dependencies between features, indicates the relative importance of members of a feature set, and helps in identifying sources of potential classification errors. Class separability yielded by different feature subsets is evaluated using the accuracy of several neural network-based classification strategies, some of them hierarchical, as well as using a feature selection technique making use of a scatter criterion. Although applied to cytogenetics, the paper presents a comprehensive, unifying methodology of qualitative and quantitative evaluation of pattern feature representation essential for accurate image classification. This methodology is applicable to many other real-world pattern recognition problems.

Accurate and high density estimation of optical flow vectors in an image sequence is accomplished by a method that estimates the velocity distribution function for small overlapping regions of the image. Because the distribution is multimodal, the method can accurately estimate the change in velocity near motion contrast borders. Large spatiotemporal support without sacrificing spatial resolution is a feature of the method, so it is not necessary to smooth the resulting flow vectors in a subsequent operation, and there is a certain degree of resistance to aperture and aliasing effects. Spatial support also provides for the accurate estimation of long-range displacements, and subpixel accuracy is achieved by a simple maximum likelihood estimation of the velocity distribution function. A confidence value is also estimated, which can help to predict flow ambiguities. The method is demonstrated using image sequences obtained from the analysis of ceramic and metal materials under stress. 1

Fast and accurate analysis of fluorescence in-situ hybridisation (FISH) images depends upon two components: a classifier to discriminate between artefacts and valid signal data, and well-discriminating features to represent the signals. After processing the image, we evaluate candidate feature sets by illustrating the probability density functions and scatter plots for the features. This analysis indicates the relative importance of members of a feature set, helps in identifying sources of potential classification errors and recommends several size, intensity and hue-based features for representing FISH signals. The recommendation is assessed by the probability of misclassification of a neural network based hierarchical strategy, and also by a feature selection technique making use of a class separability criterion. Represented by these features, nearly 90 % of valid signals and artefacts within a set of 400 test images are correctly classified. 1

We present a holistic technique for recognising words written in cursive Arabic script that does not rely on character segmentation. Each word is transformed into a normalised polar image, and a two dimensional Fourier transform is applied to the polar image. The resultant spectrum tolerates variations in size, rotation or displacement. Each word is represented by a single template, and the recognition is based on the Euclidean distance from those templates. Words are written in four different Arabic type-faces, where ligatures and overlaps pose challenges to segmentation-based methods.

We present a new technique for recognising Arabic cursive words from scanned images of text. The approach is segmentation-free, and is applied to four different Arabic typeface, where ligatures and overlaps pose challenges to segmentation-based methods. We first transform each word into a normalised polar image, then we apply a two dimensional Fourier transform to the polar image. The resultant spectrum tolerates variations in size and rotation of displacement. Each word is represented by a template that includes a set of Fourier coefficients. The recognition is based on a normalised Euclidean distance from those templates.

We present a technique for extracting structural features from cursive Arabic script. After preprocessing, the skeleton of the binary word image is decomposed into a number of segments in a certain order. Each segment is transformed into a feature vector. The target features are the curvature of the segment, its length relative to other segment lengths of the same word, the position of the segment relative to the centroid of the skeleton, and detailed description of curved segments. The result of this method is used to train the Hidden Markov Model to perform the recognition. 1

Many researchers have proposed using multiple communicating processors with a globally addressed memory for executing logic programs. However, the memory and communication contention issues characteristic of such designs have been neglected or palliated. In this paper we discuss a mew method for executing logic programs with low overhead on multiple processors. This method does not involve the sharing of memory or the copying of computation state between processors. We discuss the advantages of this approach over previous ones, and report progress on a multiprocessor implementation.

We introduce a technique, called specialisation, for automatically performing a certain class of optimisation in hierarchically specified designs composed of standard modules. Modules are subject to specialisation when they provide facilities that are not used in a design. This situation is not uncommon in ‘gate array’ and ‘standard cell’ designs used in VLSI circuits. New definitions of modules are collected in a library as they are constructed so that detailed processing need not be repeated when the same specialisation is required elsewhere in a design. The technique has been implemented in Prolog. For explanatory purposes the technique has been applied here to specialising ‘standard cell’ specifications, but the technique is generally applicable for rewriting arbitrary specifications written as collections of Horn clauses.

Abstract All clauses comprising a Prolog program are stored in a database from which they can be removed later. Other long‐term data structures are represented as clauses and are also stored and removed from the same database. Implementation techniques for the manipulation of clauses are not well known, and a lack of information has led to incorrect and incomplete implementations. Further previously unresolved issues are apparent when considering the storage of compiled clauses. We describe the way database manipulations are performed in Prolog‐X, a new compiler‐based Prolog system. We also introduce a new technique for storing the source form of compiled clauses.