Papers by Epaminondas Kapetanios
Are SKOS concept schemas ready for multilingual applications
ACM eBooks, Sep 1, 2012
This article describes our approach to accessing Knowledge Organization Systems expressed using t... more This article describes our approach to accessing Knowledge Organization Systems expressed using the Simple Knowledge Organization System (SKOS) data model. We share the view that the Web is becoming a multilingual lexical resource and a distribution infrastructure for knowledge resources. We aim to tap into this for the particular use case of Cross-Language Information Retrieval systems. The SKOS framework allows the description of monolingual or multilingual thesauri, controlled vocabularies and other classification systems in a simple machine-understandable representation. It has support for decentralized distribution on the Web of any resource described with it and includes mechanisms to interconnect different concept schemes. Yet, when building our prototype CLIR system different processes require more than the existing content of a SKOS resource: concept descriptions, labels and basic inter-concept relations. For example the SKOS concept indexing phase entails identifying potential occurrences of a SKOS concept in a text and to disambiguate based on the semantics referenced to in the overall SKOS scheme. By design, the SKOS data model does not formally define semantics of its concepts thus we have built a set of three algorithms that help generate a multilingual dataset linking to the original SKOS dataset and providing more details about the lexical entities that describe concepts. This new dataset contains specific RDF triples that aid concept identification, disambiguation and translation in CLIR.
Der in der objektorientierten Datenbank modellierte Objektraum erfaßt auch die Wissensbasis, wora... more Der in der objektorientierten Datenbank modellierte Objektraum erfaßt auch die Wissensbasis, worauf die Ableitungshistorie beruht. Dieses wird durch die Kopplung des Datenund des Prozessmodells erreicht. Auf dieser Basis sollte ein Inferenzmechanismus aufgebaut werden, damit die Konsequenzen der Wiederverarbeitungsanforderungen auf das Datenmodell ermittelt werden können.
At the cross-road of the social Web and cross-lingual information retrieval
Information Science Reference eBooks, 2009
Tanase, Diana and Kapetanios, Epaminondas (2009) At the cross-road of the social Web and cross-li... more Tanase, Diana and Kapetanios, Epaminondas (2009) At the cross-road of the social Web and cross-lingual information retrieval. In: Murugesan, San,(ed.) Handbook of research on Web 2.0, 3.0, and X. 0: technologies, business, and social applications. Information Science Reference, USA. ISBN 9781605663845
Extracting and providing knowledge within an object-centered scientific information system for atmospheric research
Abstract The paper deals with an object-centered scientific information system for the extraction... more Abstract The paper deals with an object-centered scientific information system for the extraction and preparation of scientific knowledge. A knowledge representation system (KRS) based on description logics (DL) is considered to be the basis for the construction of a terminological knowledge base. It is based on a conceptual description schema defined for the metadata needed in order to provide explanations and/or justifications to scientific results. The central element of the knowledge base is the concept and not the rule. Rules ...
IGI Global eBooks, Apr 5, 2012
Springer eBooks, 2022
Scene classification carry out an imperative accountability in the current emerging field of auto... more Scene classification carry out an imperative accountability in the current emerging field of automation. Traditional classification methods endure with tedious processing techniques. With the advent of CNN and deep learning models have greatly accelerated the job of scene classification. In our paper we have considered an area of application where the deep learning can be used to assist in the civil and military applications and aid in navigation. Current image classifications concentrate on the various available labeled datasets of various images. This work concentrates on classification of few scenes that contain pictures of people and places that are affected in the areas of flood. This aims at assisting the rescue officials at the need of natural calamities, disasters, military attacks etc. Proposed work explains a classifying system which can categorize the small scene dataset using transfer learning approach. We collected the pictures of scenes from sites and created a small dataset with different flood affected activities. We have utilized transfer learning model, RESNET in our proposed work which showed an accuracy of 88.88% for ResNet50 and 91.04% for ResNet101 and endow with a faster and economical revelation for the application involved.
A Semantically Advanced Querying Methodology for Medical Knowledge and Decision Support
IGI Global eBooks, Jan 18, 2011
A large part of all activities in healthcare deals with decision making regarding which examinati... more A large part of all activities in healthcare deals with decision making regarding which examinations and tests need to be done or, on the basis of earlier examinations, which further tests need to be ordered. Recently, guidelines for an appropriateness and necessity indication of medical interventions have been elaborated and consulted in order to evaluate the quality of decisions in specific medical domains such as cardiology and hysterectomy.
Hybrid deep convolutional neural models for iris image recognition
Multimedia Tools and Applications, Sep 12, 2021
This paper briefly explains about the application of deep learning-based methods for biometric ap... more This paper briefly explains about the application of deep learning-based methods for biometric applications. This work attempts to solve the problem of limited availability of datasets which affects accuracy of the classifiers. This paper explores the iris recognition problem using a basic convolutional neural network model and hybrid deep learning models. The augmentations used to populate the dataset and their outputs are also shown in this study. An illustration of learned weights and the outputs of intermediary stages the network like convolution layer, normalization layer and activation layer are given to help better understanding of the process. The performance of the network is studied using accuracy and receiver operating characteristic curve. The empirical results of our experiments show that Adam based optimization is good at learning iris features using deep learning. Moreover, the hybrid deep learning network with SVM performs better in iris recognition with a maximum accuracy of 97.8%. These experiments have also revealed that not all hybrid networks will give better performance as the hybrid deep learning network with KNN has given lesser accuracy.
Language Resources and Evaluation, May 1, 2020
Real-time hand movement trajectory tracking based on machine learning approaches may assist the e... more Real-time hand movement trajectory tracking based on machine learning approaches may assist the early identification of dementia in ageing deaf individuals who are users of British Sign Language (BSL), since there are few clinicians with appropriate communication skills, and a shortage of sign language interpreters. In this paper, we introduce an automatic dementia screening system for ageing Deaf signers of BSL, using a Convolutional Neural Network (CNN) to analyse the sign space envelope and facial expression of BSL signers recorded in normal 2D videos from the BSL corpus. Our approach involves the introduction of a sub-network (the multi-modal feature extractor) which includes an accurate real-time hand trajectory tracking model and a real-time landmark facial motion analysis model. The experiments show the effectiveness of our deep learning based approach in terms of sign space tracking, facial motion tracking and early stage dementia performance assessment tasks.
Proceedings of the 13th international conference on Natural Language and Information Systems: Applications of Natural Language to Information Systems
Proceedings of the International Conference on Management of Emergent Digital EcoSystems, Lyon, France, October 27 - 30, 2009
Springer eBooks, 2022
Heart rate (HR) is one of the important vital parameters of the human body and understanding this... more Heart rate (HR) is one of the important vital parameters of the human body and understanding this vital sign provides key insights into human wellness. Imaging photoplethysmography (iPPG) allows HR detection from video recordings and its unbeatable compliance over the state of art methods has made much attention among researchers. Since it is a camera-based technique, measurement accuracy depends on the quality of input images. In this paper, we present a pipeline for efficient measurement of HR that includes a learning-based super-resolution preprocessing step. This preprocessing image enhancement step has shown promising results on low-resolution input images and works better on iPPG algorithms. The experimental results verified the reliability of this method.
Lecture Notes in Computer Science, 2022
An abnormal growth develop in female uterus is uterus fibroids. Sometimes these fibroids may caus... more An abnormal growth develop in female uterus is uterus fibroids. Sometimes these fibroids may cause severe problems like miscarriage. If this fibroids are not detected it ultimately grows in size and numbers. Among different image modalities, ultrasound is more efficient to detect uterus fibroids. This paper proposes a model in deep learning for fibroid detection with many advantages. The proposed deep learning model overpowers the drawbacks of the existing methodologies of fibroid detection in all stages like noise removal, contrast enhancement, Classification. The preprocessed image is classified into two classes of data: fibroid and non-fibroid, which is done using the MBF-CDNN method. The method is validated using the parameters Sensitivity, specificity, accuracy, precision, F-measure. It is found that the sensitivity is 94.44%, specificity 95 % and accuracy 94.736%.
Ontology learning from text
ACM Computing Surveys, 2012
Ontologies are often viewed as the answer to the need for interoperable semantics in modern infor... more Ontologies are often viewed as the answer to the need for interoperable semantics in modern information systems. The explosion of textual information on the Read/Write Web coupled with the increasing demand for ontologies to power the Semantic Web have made (semi-)automatic ontology learning from text a very promising research area. This together with the advanced state in related areas, such as natural language processing, have fueled research into ontology learning over the past decade. This survey looks at how far we have come since the turn of the millennium and discusses the remaining challenges that will define the research directions in this area in the near future.
Entangled Semantics
Lecture Notes in Computer Science, 2013
In the context of monolingual and bilingual retrieval, Simple Knowledge Organisation System (SKOS... more In the context of monolingual and bilingual retrieval, Simple Knowledge Organisation System (SKOS) datasets can play a dual role as knowledge bases for semantic annotations and as language-independent resources for translation. With no existing track of formal evaluations of these aspects for datasets in SKOS format, we describe a case study on the usage of the Thesaurus for the Social Sciences in SKOS format for a retrieval setup based on the CLEF 2004-2006 Domain-Specific Track topics, documents and relevance assessments. Results showed a mixed picture with significant system-level improvements in terms of mean average precision in the bilingual runs. Our experiments set a new and improved baseline for using SKOS-based datasets with the GIRT collection and are an example of component-based evaluation
Computer Vision – ECCV 2020 Workshops
The ageing population trend is correlated with an increased prevalence of acquired cognitive impa... more The ageing population trend is correlated with an increased prevalence of acquired cognitive impairments such as dementia. Although there is no cure for dementia, a timely diagnosis helps in obtaining necessary support and appropriate medication. Researchers are working urgently to develop effective technological tools that can help doctors undertake early identification of cognitive disorder. In particular, screening for dementia in ageing Deaf signers of British Sign Language (BSL) poses additional challenges as the diagnostic process is bound up with conditions such as quality and availability of interpreters, as well as appropriate questionnaires and cognitive tests. On the other hand, deep learning based approaches for image and video analysis and understanding are promising, particularly the adoption of Convolutional Neural Network (CNN), which require large amounts of training data. In this paper, however, we demonstrate novelty in the following way: a) a multi-modal machine learning based automatic recognition toolkit for early stages of dementia among BSL users in that features from several parts of the body contributing to the sign envelope, e.g., hand-arm movements and facial expressions, are combined, b) universality in that it is possible to apply our technique to users of any sign language, since it is language independent, c) given the trade-off between complexity and accuracy of machine learning (ML) prediction models as well as the limited amount of training and testing data being available, we show that our approach is not over-fitted and has the potential to scale up.
ArXiv, 2020
Given the recent advances and progress in Natural Language Processing (NLP), extraction of semant... more Given the recent advances and progress in Natural Language Processing (NLP), extraction of semantic relationships has been at the top of the research agenda in the last few years. This work has been mainly motivated by the fact that building knowledge graphs (KG) and bases (KB), as a key ingredient of intelligent applications, is a never-ending challenge, since new knowledge needs to be harvested while old knowledge needs to be revised. Currently, approaches towards relation extraction from text are dominated by neural models practicing some sort of distant (weak) supervision in machine learning from large corpora, with or without consulting external knowledge sources. In this paper, we empirically study and explore the potential of a novel idea of using classical semantic spaces and models, e.g., Word Embedding, generated for extracting word association, in conjunction with relation extraction approaches. The goal is to use these word association models to reinforce current relatio...
Pattern Analysis and Applications, 2019
Physical traits such as the shape of the hand and face can be used for human recognition and iden... more Physical traits such as the shape of the hand and face can be used for human recognition and identification in video surveillance systems and in biometric authentication smart card systems, as well as in personal health care. However, the accuracy of such systems suffers from illumination changes, unpredictability, and variability in appearance (e.g. occluded faces or hands, cluttered backgrounds, etc.). This work evaluates different statistical and chrominance models in different environments with increasingly cluttered backgrounds where changes in lighting are common and with no occlusions applied, in order to get a reliable neural network reconstruction of faces and hands, without taking into account the structural and temporal kinematics of the hands. First a statistical model is used for skin colour segmentation to roughly locate hands and faces. Then a neural network is used to reconstruct in 3D the hands and faces. For the filtering and the reconstruction we have used the growing neural gas algorithm which can preserve the topology of an object without restarting the learning process. Experiments conducted on our own database but also on four benchmark databases (Stirling's, Alicante, Essex, and Stegmann's) and on deaf individuals from normal 2D videos are freely available on the BSL signbank dataset. Results demonstrate the validity of our system to solve problems of face and hand segmentation and reconstruction under different environmental conditions.
Uploads
Papers by Epaminondas Kapetanios