Although several automatic computer systems have been proposed to address facial expression recognition problems, the majority of them still fail to cope with some requirements of many practical application scenarios. In this paper, one of the most influential and common issues raised in practical application scenarios when applying automatic facial expression recognition system, head pose variation, is comprehensively explored and investigated. In order to do this, two novel texture feature representations are proposed for implementing multi-view facial expression recognition systems in practical environments. These representations combine the block-based techniques with Local Ternary Pattern-based features, providing a more informative and efficient feature representation of the facial images. In addition, an in-house multi-view facial expression database has been designed and collected to allow us to conduct a detailed research study of the effect of out-of-plane pose angles on the performance of a multi-view facial expression recognition system. Along with the proposed in-house dataset, the proposed system is tested on two well-known facial expression databases, CK+ and BU-3DFE datasets. The obtained results shows that the proposed system outperforms current state-of-the-art 2D facial expression systems in the presence of pose variations.

In the context of facial expression recognition (FER), this paper reviews the fundamental theories of emotions and further explains the key dimensions of a defined emotional space. The main contribution of this paper is to propose a set of novel categorization methods for facial expressions to be used in the design of an automatic FER system. This novel categorization enables the facial expression to be interpreted in a better way that and to be more effective in practical applications of automatic FER systems. In order to validate the feasibility of the proposed categorization methods, a set of experiments is reported which investigates and analyzes the influence that the novel categorization brings to a multi-view FER system.

Digital palaeography is an emerging research area which aims to introduce digital image processing techniques into palaeographic analysis for the purpose of providing objective quantitative measurements. This paper explores the use of a fully automated handwriting feature extraction, visualization, and analysis system for digital palaeography which bridges the gap between traditional and digital palaeography in terms of the deployment of feature extraction techniques and handwriting metrics. We propose the application of a set of features, more closely related to conventional palaeographic assesment metrics than those commonly adopted in automatic writer identification. These features are emprically tested on two datasets in order to assess their effectiveness for automatic writer identification and aid attribution of individual handwriting characteristics in historical manuscripts. Finally, we introduce tools to support visualization of the extracted features in a comparative way, showing how they can best be exploited in the implementation of a content-based image retrieval (CBIR) system for digital archiving.

As biometric systems are deployed in increasingly diverse applications, it becomes correspondingly important to understand the impact which human aging has on system performance. Aging directly affects those physiological and behavioral traits which are characterized in biometric measurements, and a practical biometric system must be designed to account for age-induced changes. However, age can also have very positive implications, for example as a source of further identification information. This paper reviews research to understand how age factors impinge on biometric systems and uses this to synthesize a system infrastructure to unify implementation principles. We present new results to show how multiagent structures can provide an effective framework for this purpose, enhancing performance in both identification and predictive scenarios.

Clustering is deemed one of the most difficult and challenging problems in machine learning. In this paper, we propose a multilocal search and adaptive niching-based genetic algorithm with a consensus criterion for automatic data clustering. The proposed algorithm employs three local searches of different features in a sophisticated manner to efficiently exploit the decision space. Furthermore, we develop an adaptive niching method, which can dynamically adjust its parameter value depending on the problem instance as well as the search progress, and incorporate it into the proposed algorithm. The adaptation strategy is based on a newly devised population diversity index, which can be used to promote both genetic diversity and fitness. Consequently, diverged niches of high fitness can be formed and maintained in the population, making the approach well-suited to effective exploration of the complex decision space of clustering problems. The resulting algorithm has been used to optimize a consensus clustering criterion, which is suggested with the purpose of achieving reliable solutions. To evaluate the proposed algorithm, we have conducted a series of experiments on both synthetic and real data and compared it with other reported methods. The results show that our proposed algorithm can achieve superior performance, outperforming related methods.

Biometric-based identification systems can offer several advantages over traditional forms of identity authentication. However, concerns have been raised about the privacy of the personal biometric data, since these systems need to ensure their integrity and public acceptance. In order to address these issues, the notion of cancellable biometrics was introduced. It describes biometric templates that can be cancelled and replaced, in case of being lost or stolen. However, this concept still raises new issues, since they make the authentication problem more complex and difficult to solve. Thus, more effective authentication structures are needed to perform these tasks. In this paper, we investigate the use of ensemble systems in cancellable biometrics, using online signature identification. In order to improve the effectiveness of the ensemble systems, we used genetic algorithms in the choice of an optimized set of weights that are used along with the output of the individual classifiers to define the final output of the system. In addition, we proposed the use of genetic algorithm in the procedure to create the cancellable biometric data, aiming to obtain more efficient cancellable data. The main of this paper is to provide more security in the biometric-based identification process.

In this paper, we describe DocExplore, an integrated software suite centered on the handling of digitized documents with an emphasis on ancient manuscripts. This software suite allows the augmentation and exploration of ancient documents of cultural interest. Specialists can add textual and multimedia data and metadata to digitized documents through a graphical interface that does not require technical knowledge. They are helped in this endeavor by sophisticated document analysis tools that allows for instance to spot words or patterns in images of documents. The suite is intended to ease considerably the process of bringing locked away historical materials to the attention of the general public by covering all the steps from managing a digital collection to creating interactive presentations suited for cultural exhibitions. Its genesis and sustained development reside in a collaboration of archivists, historians and computer scientists, the latter being not only in charge of the development of the software, but also of creating and incorporating novel pattern recognition for document analysis techniques.

Automated identification of individuals using biometric technologies is finding increasing application in diverse areas, yet designing practical systems can still present significant challenges. Choice of the modality to adopt, the classification/matching techniques best suited to the application, the most effective sensors to use, and so on, are all important considerations, and can help to ameliorate factors which might detract from optimal performance. Less well researched, however, is how to optimise performance by means of exploiting broader-based information often available in a specific task and, in particular, the exploitation of so-called "soft" biometric data is often overlooked. This paper proposes a novel approach to the integration of soft biometric data into an effective processing structure for an identification task by adopting a fuzzy representation of information which is inherently continuous, using subject age as a typical example. Our results show this to be a promising methodology with possible benefits in a number of potentially difficult practical scenarios.

In the field of cancellable (revocable) biometrics much attention has been paid to physiological biometrics(e.g. fingerprint, iris) when compromised. However, the potential for natural revocability in behavioural biometrics(e.g. handwritten signature) is often overlooked, and the lack of databases to support this type of investigation imposes limitations. This paper presents a preliminary study of natural revocability, revealing relationships among established and new signature forms in relation to achievable stability. We study also the implications at the classification level.

A novel approach to the problem of keyword retrieval in cursive handwritten documents is introduced in this work. Two issues are addressed: small dataset size and uneven sample distribution across the character set. The proposed strategies utilise graphemes (fragments of a handwritten word) to implement a recognition model which is subsequently used to form the feature model for the query word.

Purpose Biometric authentication, which requires storage of biometric templates and/or encryption keys, raises a matter of serious concern, since the compromise of templates or keys necessarily compromises the information secured by those keys. To address such concerns, efforts based on dynamic key generation directly from the biometrics have recently emerged. However, previous methods often have quite unacceptable authentication performance and/or small key spaces and therefore are not viable in practice. The purpose of this paper is to propose a novel method which can reliably generate long keys while requires storage of neither biometric templates nor encryption keys. Design/methodology/approach This proposition is achieved by devising the use of fingerprint orientation fields for key generation. Additionally, the keys produced are not permanently linked to the orientation fields, hence, allowing them to be replaced in the event of key compromise. Findings The evaluation demonstrates that the proposed method for dynamic key generation can offer both good reliability and security in practice, and outperforms other related methods. Originality/value In this paper, the authors propose a novel method which can reliably generate long keys while requires storage of neither biometric templates nor encryption keys. This is achieved by devising the use of fingerprint orientation fields for key generation. Additionally, the keys produced are not permanently linked to the orientation fields, hence, allowing them to be replaced in the event of key compromise.

The effectiveness with which individual identity can be predicted in, for example, an antiterrorist scenario can benefit from seeking a broad base of identity evidence. The issue of improving performance can be addressed in a number of ways, but system configurations based on integrating different information sources (often involving more than one biometric modality) are a widely adopted means of achieving this. This paper presents a new approach to improving identification performance, where both direct biometric samples and “soft-biometric” knowledge are combined. Specifically, however, we propose a strategy based on an intelligent agent-based decision-making process, which predicts both absolute identity and also other individual characteristics from biometric samples, as a basis for a more refined and enhanced overall identification decision based on flexible negotiation among class-related agents.

In this paper, we present a novel and robust road tracking system for vision-based personal navigation. Novelty of the work includes the use of multiple Condensation filters to track the road of arbitrary shape and automatic switching between trackers according to road conditions. The approach allows the road to be represented as a simple hyperbola. It also supports the representation of the road as a sequence of connected arcs/segments so that information from a digital map can be integrated into tracking. The parameters of the hyperbola road model are estimated using multiple vanishing points located in image strips. The road tracking method is robust in dealing with complex road shapes, background clutters, shadows, and road markings. Experiments using real videos demonstrate the robustness of our approach.

A new multimodal biometric database designed and acquired within the framework of the European BioSecure Network of Excellence is presented. It is comprised of more than 600 individuals acquired simultaneously in three scenarios: 1) over the Internet, 2) in an office environment with desktop PC, and 3) in indoor/outdoor environments with mobile portable hardware. The three scenarios include a common part of audio/video data. Also, signature and fingerprint data have been acquired both with desktop PC and mobile portable hardware. Additionally, hand and iris data were acquired in the second scenario using desktop PC. Acquisition has been conducted by 11 European institutions. Additional features of the BioSecure Multimodal Database (BMDB) are: two acquisition sessions, several sensors in certain modalities, balanced gender and age distributions, multimodal realistic scenarios with simple and quick tasks per modality, cross-European diversity, availability of demographic data, and compatibility with other multimodal databases. The novel acquisition conditions of the BMDB allow us to perform new challenging research and evaluation of either monomodal or multimodal biometric systems, as in the recent BioSecure Multimodal Evaluation campaign. A description of this campaign including baseline results of individual modalities from the new database is also given. The database is expected to be available for research purposes through the BioSecure Association during 2008.

Although it has long been recognized that non-biometric factors (for example, general demographic characteristics) can have an impact on the performance of automated person identification systems, such information is not routinely adopted in most practical biometric processing. In forensic applications, however, such additional information may be exploited most productively, since typical scenarios require the prediction of a wide range of individual characteristics from a range of available samples, biometric and non-biometric. In many such situations, it may be very useful to predict characteristics short of actual identity, since this type of prediction can significantly increase the amount of evidence available. In this paper we provide some benchmarking data to demonstrate the extent to which typical non-biometric information might be effective in practice in this context, and illustrate the differential effects depending on the classifier adopted. In particular, however, we investigate experimentally how traditionally measured handwritten signature data might be used to predict non-biometric categories in a way which can be exploited a variety of practical application scenarios.

In this paper, we develop a fingerprint matching method which operates by first introducing an integrated measure, which combines two different matching criteria based on heterogeneous features. We then devise a genetically guided algorithm to optimise the integrated measure for simultaneous fingerprint alignment and verification. The proposed method is evaluated through experiments conducted on two public domain collections of fingerprint images and compared with related work, with very encouraging results.

Biometric authentication is increasingly gaining popularity in a wide range of applications. However, the storage of the biometric templates and/or encryption keys that are necessary for such applications is a matter of serious concern, as the compromise of templates or keys necessarily compromises the information secured by those keys. In this paper, we propose a novel method, which requires storage of neither biometric templates nor encryption keys, by directly generating the keys from statistical features of biometric data. An outline of the process is as follows: given biometric samples, a set of statistical features is first extracted from each sample. On each feature subset or single feature, we model the intra and interuser variation by clustering the data into natural clusters using a fuzzy genetic clustering algorithm. Based on the modelling results, we subsequently quantify the consistency of each feature subset or single feature for each user. By selecting the most consistent feature subsets and/or single features for each user individually, we generate the key reliably without compromising its relative security. The proposed method is evaluated on handwritten signature data and compared with related methods, and the results are very promising.

In this paper, we present a robust road detection and tracking method based on a condensation particle filter for real-time video-based navigation applications. The image is divided into horizontal strips, and vanishing point (VP) detection is performed on each image strip. We propose a method for estimating the density of road boundary line segments in the image so that VP detection in an image strip takes into account the detection results in the neighboring image strips. This use of contextual information for VP detection leads to more accurate detection results. The estimated road parameters are then used to initialize the condensation tracker. Experiments using real road videos demonstrate the robustness of our method to difficult road conditions due to the presence of partial occlusion, shadows, and road signs.

Clustering is inherently a difficult task, and is made even more difficult when the selection of relevant features is also an issue. In this paper we propose an approach for simultaneous clustering and feature selection using a niching memetic algorithm. Our approach (which we call NMA_CFS) makes feature selection an integral part of the global clustering search procedure and attempts to overcome the problem of identifying less promising locally optimal solutions in both clustering and feature selection, without making any a priori assumption about the number of clusters. Within the NMA_CFS procedure, a variable composite representation is devised to encode both feature selection and cluster centers with different numbers of clusters. Further, local search operations are introduced to refine feature selection and cluster centers encoded in the chromosomes. Finally, a niching method is integrated to preserve the population diversity and prevent premature convergence. In an experimental evaluation we demonstrate the effectiveness of the proposed approach and compare it with other related approaches, using both synthetic and real data.

The collection of human biometric test data for system development and evaluation within any chosen modality generally requires significant time and effort if data are to be obtained in workable quantities. To overcome this problem, techniques to generate synthetic data have been developed. This paper describes a novel technique for the automatic synthesis of human handwritten-signature images, which introduces modeled variability within the generated output based on positional variation that is naturally found within genuine source data. The synthesized data were found to generate similar verification rates to those obtained using genuine data with the use of a commercial verification engine, thereby indicating the suitability of the data synthesized by using this method for a wide range of application scenarios.

We present a a statistical approach to skew detection, where the textual features of a document image are modeled as a mixture of straight lines in Gaussian noise. The EM algorithm is used to estimate the parameters of the mixture model and the skew angle estimate is extracted from the estimated parameters. Experiments prove that our method has some advantages over other existing methods in terms of accuracy and efficiency.

This paper describes a novel method for the generation of synthetic handwritten signatures, in the form of a series of time-stamped pen data channels, for use in dynamic signature verification experimentation. The technique introduces modelled variability within the generated data based on variation that is naturally found within genuine source data. Experimentation using the SVC2004 dataset and a commercial signature verification engine shows that the synthesized data achieves comparative verification performance to the use of genuine data.

Performance in terms of accuracy is one of the most important goal of a biometric system. Hence, having a measure which is able to predict the performance with respect to a particular sample of interest is specially useful, and can be exploited in a number of ways. In this paper, we present two automatic measures for predicting the performance in off-line signature verification. Results obtained on a sub-corpus of the MCYT signature database confirms a relationship between the proposed measures and system error rates measured in terms of equal error rate (EER), false acceptance rate (FAR) and false rejection rate (FRR).

Minutiae point pattern matching is the most common approach for fingerprint verification. Although many minutiae point pattern matching algorithms have been proposed, reliable automatic fingerprint verification remains as a challenging problem, both with respect to recovering the optimal alignment and the construction of an adequate matching function. In this paper, we develop a memetic fingerprint matching algorithm (MFMA) which aims to identify the optimal or near optimal global matching between two minutiae sets. Within the MFMA, we first introduce an efficient matching operation to produce an initial population of local alignment configurations by examining local features of minutiae. Then, we devise a hybrid evolutionary procedure by combining the use of the global search functionality of a genetic algorithm with a local improvement operator to search for the optimal or near optimal global alignment. Finally, we define a reliable matching function for fitness computation. The proposed algorithm was evaluated by means of a series of experiments conducted on the FVC2002 database and compared with previous work. Experimental results confirm that the MFMA is an effective and practical matching algorithm for fingerprint verification. The algorithm is faster and more accurate than a traditional genetic-algorithm-based method. It is also more accurate than a number of other methods implemented for comparison, though our method generally requires more computational time in performing fingerprint matching.

The reported work aims to objectively and accurately assess the post-stroke clinical condition of visuo-spatial neglect using a series of standardized geometric shape drawing tasks. We present a method implementing existing pencil-and-paper diagnostic methods and define a set of static and dynamic features that can be extracted from drawing responses captured online using a graphics tablet. We also present a method for automatically assessing the constructional sequence of the drawing using Hidden Markov Models. The method enables the automated extraction, position identification and drawing order of individual sides of a shape within a drawing. Discrimination between two populations (a neglect population and stroke subjects without neglect as determined by existing standard assessment methods) using a combination of performance features and constructional sequence is examined across three separate drawing tasks. Results from experimentation show how a combination of sequence and performance features is able to generalize across a wide variety of input samples and obtain a diagnostic classification which can be used alongside other forms of conventional assessment. Furthermore, the application of a multi-classifier combination strategy leads to a significant increase in recognition ability.

Document security is an increasingly importantelement in the multi-faceted discipline ofdocument processing, and authentication ofindividual identity will play an increasinglyimportant future role in relation to questions ofdocument ownership, identity andconfidentiality. Biometrics-based techniques areemerging as key elements in the drive to addresssecurity and confidentiality in an effective way,yet past experience suggests that there are manypractical issues yet to be resolved if biometrictechnologies are to fulfill their potential in thedocument processing field. This paper addressessome aspects of biometric processing which arebecoming increasing priorities, and suggestshow a greater engagement of the documentprocessing community can help to bring aboutrefinements to existing approaches to biometricidentity checking.

A method for automatically assessing the constructional sequence from a neuropsychological drawing task using Hidden Markov Models is presented. We also present a method of extracting and identifying the position of individual pen strokes relating to individual sides of a shape within a drawing to form training and testing sequences. Our results from two experiments using data from patients with visuo-spatial neglect show the HMM classifier is able to generalise on incorrectly extracted sequences and obtain a diagnostic classification which can be used alongside other forms of conventional assessment. 1.

In this paper initially we propose a novel approach toclassify handwritten characters based on a directional decompositionof the corresponding chain-code representation.This is alternative to previous transformations of thechain-codes proposed by the authors, namely the orderedand random decomposition of the bit-planes resulting fromthe binary representation of the chain-codes. Subsequentlywe utilize the power of the recently developed multiple classifierschemes using sntuple classifiers to integrate the complimentaryinformation encapsulated in all three transformationsinto a more powerful and robust character recognitionsystem. The results obtained through a series ofcross-validation experiments show that the proposed fusionscheme not only outperforms its constituent parts and anumber of other successful classifiers, but also enables significantsavings in memory requirements compared to theoriginal sntuple-based recognition system.

A novel strategy for the representation and manipulationof distributed documents, potentially complex andheterogeneous, is presented in this paper. The documentunder the proposed model is represented in a hierarchicalstructure. Associated metadata' describes the flexiblehierarchy with the scope of dynamically restructuring thetree at runtime. All useful functionals can also be includedwithin the hierarchy to minimize reliance on externalprograms in manipulating sensitive data. Thisgives the proposed model two key properties: generality(capable of representing any document format includingfuture innovations) and autonomy (non-reliance on externalprograms). The model also allows incorporation ofadditional features for security and access control. Biometricperson authentication measures are introduced. Abrief example illustrates the key ideas.

This paper investigates the performance of multi-neural systems, focusing on the benefits that can be gained when integrating different types of neural experts (hybrid multi-neural system). An empirical evaluation shows that the integration of different types of neural networks leads to an improvement in performance in a practical classification task for a range of combination methods.

The set of fuzzy connectives can be seen as an important combination tool, such as in combining the antecedent sets of the rules, in multi-criteria decision making and in combining the outputs of neural classifiers in a multi-neural system. This papers investigates the performance of some fuzzy combination schemes applied to a multi hybrid neural system which is composed of neural and fuzzy neural networks. An empirical evaluation in a handwritten numeral recognition task is used to investigate the performance of the presented fuzzy methods with some existing combination methods.

This paper examines a computer-based technique for the detection of visuo-spatial neglect from the responses of a simple geometric shape copying task. Defining pass/fail criteria based on the presence of drawn components, responses can be accurately and objectively assessed. More importantly, we show that by analysing novel dynamic performance features detailing timing and constructional aspects of each response, significant performance deficits can be noted in drawings made by clinically diagnosed neglect subjects that would have been classified as 'normal' using conventional static analysis, thus improving the sensitivity of the assessment.

The moving window classifier (MWC) is a simple and efficient classifier structure which, although shown to be capable of promising performance in a variety of tasks such as face recognition, its common application is a tool in text recognition. Various measures have been proposed to improve the MWC classification speed and to reduce memory space requirement. This paper introduces techniques for improving the MWC classification accuracy without losing any of gains previously achieved. These performance enhancement schemes are readily applicable to a range of related classifiers and hence provide a generalized method for enhancement in a variety of tasks.

A novel strategy for the representation and manipulation of multimedia documents is introduced. The model allows such documents to be treated as autonomous entities capable of representing and modifying their own internal structures. The model is sufficiently flexible as to allow any given arbitrary document format and any programming language to be incorporated within its implementation.

In this paper we introduce a global optimisation technique, namely a genetic algorithm, into a parallel multiclassifier system design process. As few similar systems have been proposed to date our main focus in this study is to explore the statistical properties of the self-configuration process in order to enhance our understanding of its internal operational mechanism and to propose possible improvements. For this we tested our system in a series of character recognition tasks ranging from printed to handwritten data. Subsequently, we compare its performance with that of two alternative multiple classifier combination strategies. Finally, we investigate, over a set of cross-validating experiments, the relation between the performances of the individual classifiers and their variability, and the frequency with which each of them is chosen to participate in the final configuration generated by the genetic algorithm.

We introduce a novel multiple classifier system with the ability of automatic self-configuration. The system employs a genetic algorithm to optimise the configuration of the participating individual classifiers arranged in a parallel structure. Our primary interest was to study the behaviour of such an integrated system, first in the case of increasingly complex tasks and secondly when additional information is made available in the form of increasingly larger training data sets. The fact that these cases often arise in real world applications underline their special importance in developing classifier systems that can address realistic problem domains. As an example we tested the proposed system on a character recognition task using one printed, and one handwritten, character data set. Our findings strongly suggest that significant benefit can be gained from the integration of the genetic algorithm-based optimisation process into the system in both situations.

This paper presents a mechanism to vary the vigilance parameter in the RePART fuzzy neural network. This mechanism helps to smooth out the problem of category proliferation which affects ARTMAP-based networks. Empirical experiments show that the use of variable vigilance improves the performance of the RePART model while, at the same time, requiring a less complex structure.

A novel multiple-expert framework for recognition of handwritten characters is presented. The proposed framework is composed of multiple classifiers (experts) put together in such a manner as to enhance the recognition capability of the combined network compared to the best performing individual expert participating in the framework. Each of these experts has been derived from a novel neural structure in which the weight values are derived from Clifford algebra. A Clifford algebra is a mathematical paradigm capable of capturing the interdimensional dependencies found in multidimensional data. It offers a technique for concise data storage and processing by representing dependencies between the component dimensions of the data which is otherwise difficult to encode and hence is often employed in analyzing multidimensional data. Results achieved by the proposed multiple-expert framework demonstrates significant improvement over alternative techniques.

The Moving Window Classifier(MWC) has previously been proposed as an efficient scheme for text recognition applications. In this paper, the potential of the MWC algorithm in face recognition is investigated. To maintain the memory requirements of the classifier within acceptable practical limits, the concept of bit-plane encoding is utilized. The experimental results reported show very encouraging performance for both the schemes. 1

Visuo-spatial neglect (VSN) is a condition following a stroke or head injury whereby a patient fails to respond to stimuli on one side of the visual field. A standard clinical assessment technique for analysis of VSN is a pencil-and-paper based figure drawing task. Traditional static analysis of this task involves assessing the presence of the major components of the drawing. Marking of drawings is subjective, relying on assessors' own judgement and experience, and therefore no standardisation exists between assessors. Using a computer-based test capture system, this paper establishes a standardised performance assessment for a drawing task including a series of novel dynamic performance features pertaining to the timing and constructional aspects of test performance. A case sturdy of two patients demonstrates the ability to detect VSN from a response which would have traditionally been assessed as normal and hence improve the sensitivity of the task.

Implementation of various multiple expert decision combination approaches for recognition of machine printed characters involves costs in terms of number of processing elements and the overall layout of the decision fusion hierarchy. The paper investigates the relationship between the implementation cost and the performance achieved by these approaches. Seven different multiple expert decision combination approaches have been investigated in terms of this cost versus performance profile.

In many applications of computer vision, combination of decisions from multiple sources is a very important way of achieving more accurate and robust classification. Many such techniques can be used, two of which are the Majority Voting and the Divide and Conquer techniques. The former achieves decision combination by measuring consensus among the participating classifiers and the latter achieves the same by dividing the problem into smaller problems and solving each of these sub-problems more efficiently. Both these approaches have their advantages and disadvantages. In this paper, a novel approach to combining these two techniques is presented. Although the success of the approach has been demonstrated in a typical application area of computer vision (recognition of complex and highly variable image data), the approach is completely generalised and is applicable to other task domains.

It is well known that substantial improvements can be obtained in difficult pattern recognition problems by combining or integrating the outputs of multiple neural classifiers. This paper analyses the performance of some combination schemes applied to a multi-hybrid neural system which is composed of neural and fuzzy neural networks. Essentially, the combination methods employ different ways to extract valuable information from the output of the experts through the use of confidence (weights) measures of the ensemble members to each class. An empirical evaluation in a handwritten numeral recognition task is used to investigate the performance of the presented methods in comparison with some existing combination methods.

This paper introduces a novel networking strategy for RAM-based neurons which significantly improves the training and recognition performance of such networks whilst maintaining the generalization capabilities achieved in previous network configurations. A number of different architectures are introduced, each using the same underlying principles. Initially, features which are common to all architectures are described, illustrating the basis of the underlying paradigm. Three architectures are then introduced illustrating different techniques of employing the paradigm to meet differing performance specifications. The architectures are described in terms of the structure of the neurons they employ. Details of the various training and recognition algorithms employed by the architectures are supplied in order to present a complete description of the operation of this class of artificial neural network.

A novel confidence-based parallel multiple expert decision combination framework is introduced. The traditional approaches to parallel multiple source decision fusion either take no account of the confidence of each deci-sion by each participating expert in the combined framework or only exploit the confidences associated with each decision. But it is entirely possible to incorporate more additional a priori information in the form of various confidence indices that can be estimated from the performances of various participating experts. The confidence-based parallel multiple expert decision combination framework proposed here addresses this shortcoming. Very en-couraging results have been obtained by implementing this proposed frame-work in combining decisions of multiple experts applied to the problem of handwritten and machine printed character recognition. 1

A new topology for classifying decision combinations of multiple experts in the framework of a multiple expert character recognition platform is introduced. It is demonstrated that many existing multiple expert configurations for character recognition can be categorised by using this method of defining classification strategies. It is also demonstrated that the design of multiple expert character recognition configurations can be streamlined by classifying these structures in terms of how the channels used for carrying information among different experts are interconnected irrespective of the algorithms used by cooperating experts and by the final decision combination expert. Case studies of actual multiple expert character recognition configurations have been investigated and it is shown how they can be categorised with respect to the decision combination topologies introduced in the paper.

This paper presents a completely integrated Boolean neural architecture, where a selforganizing Boolean neural network (SOFT) is used as a front-end processor to a feedforward Boolean network based on goal-seeking principles (GSN f). This paper will evaluate the advantages of the integrated SOFT-GSN f over GSN f by showing its increased effectiveness in an optical character recognition task.

Dynamic Time Warping is a mathematical optimisation technique for sequentially structured problems, which has, over the years, played a major role in providing primary algorithms for automatic signature verification. As this useful method of non-linear, elastic time alignment, still has a high computational complexity due to the repetitive nature of its operations for the optimisation process, this paper proposes an algorithm using a pipeline paradigm this being chosen with the intention of overcoming possible dead-locks in the highly distributed network. The algorithm was implemented on a transputer network on the Meiko Computing Surface using Occam2, which resulted in a reduction of time complexity by an order of magnitude.

This paper describes possible strategies for the implementation of a feature selection algorithm particularly suited to the realisation of an efficient automatic handwritten signature verification system in which an active feature vector, optimised with respect to an individual signer, is constructed during an enrollment period. A range of configurations based on transputer arrays are considered and the possible implementational approaches evaluated. The paper demonstrates how the inherent parallelism which exists within a generic model for verification can be exploited to provide an optimised general-purpose framework for verification processing.

Describes a novel approach to automatic signature verification based on the optimisation of a descriptive feature vector for each individual signer. The nature of the optimisation process is such that user enrolment becomes highly computationally intensive and hence a parallel processing approach to system implementation is necessary for practical applications. The design approach is evaluated and practical results are presented.>

The existence of an upper bound for the error probability as a function of I-divergences between an original and an approximating distribution is proved. Such a bound is shown to be a monotonic nondecreasing function of the I-divergences, reaching the Bayes error probability when they vanish. It has been shown that if the closeness between the original and approximating distributions is assessed by the probability of error associated with a particular two-class recognition problem in which those functions are the class conditional distributions, then the best upper bound for such probability is ¿ regardless of the value of the I-divergences between them. Approaching the approximation problem from a rather different viewpoint, this correspondence considers the problem of a two-class discrete measurement classification where the original distributions are replaced by approximations, and its effects on the probability of error. The corresponding analysis is presented in detail.