Proceedings of the Third International Conference on Intelligent Human Computer Interaction (IHCI 2011), Prague, Czech Republic, August, 2011

While emotional intelligence plays a key role in facilitating satisfactory interactions between people, its application is both underexplored and underexploited in human–computer interaction. Self-service technology is increasingly being incorporated by goods and service providers, however user satisfaction is still less than ideal. Studies have been carried out in which an affective embodied agent has been found to reduce frustration in users of interactive computer systems. This paper presents a preliminary study as a part of research aimed at designing and implementing an agent which detects negative emotions in a human user and expresses its own emotional reaction with the aim of improving the user’s mood and therefore their level of satisfaction in the context of a self-service interaction. We describe a study to determine customer facial expressions using facial Action Units (AUs) during interactions with self-service supermarket checkouts. Our preliminary results indicate that AU 23 and AU 24 were displayed with particular frequency.

The growth of Rich Internet Applications (RIAs) calls for new conceptual tools that enable designers and web engineers to model and keep under control the design complexity unleashed by innovative interaction (with increasing communication potential) and carefully consider the impact of the design decisions on the optimal flow of the user experience. Based on the theory of Situational Awareness, in this paper we illustrate how 5 major “interface demons” are particularly relevant for RIA engineering and undermine an effective dialogue between users and RIA interfaces. From this analysis, we propose a set of conceptual design primitives (Rich-IDM) to enable designers and web engineers to characterize the complex components of RIA models and take design decisions which meet both usability and communication requirements.

People with visual impairments, particularly blind people face alot of difficulties browsing the web with assistive technologies such as screen readers, when websites do not conform to accessibility standards and are thus inaccessible. HTML is the basic language for website design but its ALT attribute on the IMG element does not adequately capture comprehensive image semantics and description in a way that can be accurately interpreted by screen readers, hence blind people do not usually get the complete description of the image. Most of the problems however arise from web designers and developers not including a description of an image or not comprehensively describing these images to people with visual impairments. In this paper, we propose the use of the Image Description Assessment Tool (IDAT), a Java-based tool containing some proposed heuristics for assessing how well an image description matches the real content of the image on the web. The tool also contains a speech interface which can enable a visually impaired individual to listen to the description of an image that has been uploaded unto the system.

With the rapid development of information technology, the emphasis on the quality of user interfaces has been increasing recently, also with regard to mobile platforms, accessibility etc. In this paper we focus on engaging more interactive ways to image search. While observing and discussing with users about how they wish to proceed during search of images we detected four typical scenarios. We present all of them on concrete examples. We also describe what kind of image features our system works with and how we detect them. We introduce our own user interface of Xingle testing system where are all the mentioned scenarios implemented. The Xingle system works with about half a million images that were collected for testing purposes.

Brain-Computer Interface is aimed as a direct communication pathway between human or animal brain and an external device. A reliable, accurate and fast identification of a being’s intention based on EEG signal scanning is crucial part of the system. To improve the classification accuracy we propose to use Independent Component Analysis for \(\mu \)-rhythm identification in data corresponding to motor imagery task performance during Brain-Computer Interface training and operation. We show that independent components corresponding to the \(\mu \)-rhythm allow for higher classification accuracy comparing to raw EEG recordings usage.

Document clustering refers to unsupervised classification (categorization) of documents into groups (clusters) in such a way that the documents in a cluster are similar, whereas dissimilar documents are assigned in different clusters. The documents may be web pages, blog posts, news articles, or other text files. A popular and computationally efficient clustering technique is flat clustering. Unlike hierarchical techniques, flat clustering algorithms aim to partition the document space into groups of similar documents. The cluster assignments however may be hard or soft. This paper presents our experimental work on evaluating some hard and soft flat-clustering algorithms, namely K-means, heuristic k-means and fuzzy C-means, for categorizing text documents. We experimented with different representations (tf, tf.idf, Boolean) and feature selection schemes (with or without stop word removal and with or without stemming) on some standard datasets. The results indicate that tf.idf representation and the use of stemming obtains better clustering. Moreover, fuzzy clustering obtains better results than K-means on almost all datasets, and is also a more stable method.

A Natural Language Interface (NLI) is studied and developed for retrieving the recorded domain information within a prototype system, which manages complex information within the paradigm of recording live interactive records of information so that the recorded information can be queried and replayed. After introducing the background of the prototype system, concrete querying examples are presented so as to demonstrate the effectiveness and efficiency of using the developed NLI. This is done by comparing the natural language queries and the corresponding SQL statements, which are automatically gained within the prototype system. This paper shows that an NLI can greatly reduce the efforts for writing or understanding queries of information systems, thus can enhance the system usability.

Natural language processing is showing more interest in the Amazigh language in recent years. Suitable resources for Amazighe are becoming a vital necessity for the progress of this research. Corpora are an important resource but Amazighe lacks sufficient resources in this field, therefore we have been conducted to build an Amazighe corpus. In this paper, we present preliminary result experiments with a corpus for Standard Amazighe. We selected samples of published data from different Amazighe varieties. The selection was driven mainly by the amount of data available. We still demonstrate the completeness and representativeness of this corpus using metrics and show its suitability for language engineering experiments.

Processing of Arabic language is eminent for the fact that currently the number of computer and Internet users in the Arab word is growing tremendously. The problem of stemming is very important in information retrieval, knowledge mining and language processing. Arabic has very complex morphology and stemming rules that must deal with many specific properties of Arabic. This paper describes very simple rules for stemming of Arabic words. Two of these rules are universal, i.e. they are applicable to any word category, and one rule for each of the four categories: nouns, verbs, adverbs and adjectives. The rules were more successful in case of adverbs. As for nouns, verbs and adjectives, some errors occurred especially in case of suffix processing.

This work describes a novel fuzzy logic system designed to meet the real world demand of providing intelligent ranking to large repositories of documents previously encoded with non-fuzzy (crisp) metadata. The fuzzy logic prototype was tested in practice to complement the GeoConnections Discovery Portal, which is a web portal for specialized search and retrieval of Canadian geographic data resources via an associated web service. Users of the portal are able to query the system and then filter their search results by selecting topic categories, spatial and temporal extents, and resource types. The authors present a fuzzy logic information retrieval system that utilizes document metadata, and compare it to an unranked listing, standard term frequency-inverse document frequency (TF-IDF) ranking, and a TF-IDF/fuzzy hybrid system. Results indicate that the fuzzy logic system provided the overall highest precision among the top ranked documents for searches by an expert user, and that these results were robust with respect to the number of results returned by a number of different query types.

Our goal is to capture style from real human motion so it can be rendered with a virtual agent that represents this human user. We used expressivity parameters to describe motion style. As a first contribution, we propose an approach to estimate a subset of expressivity parameters defined in the literature (namely spatial extent and temporal extent) from captured motion trajectories. Second, we capture the expressivity of real users and then output it to the Greta engine that animates a virtual agent representing the user. We experimentally demonstrate that expressivity can be another clue for identifiable virtual clones of real humans.

Today’s data needs flexible visualization, and the advent of new tech-nologies defined under HTML5 has opened up new possibilities for rich web-applications. We used JavaScript and the new HTML5 ‘canvas’ element, to create a software module GraphVis for online visualization and exploration of large relationship networks. As a first test-case we applied the module to Princeton’s WordNet English thesaurus, and made the result available as the web-application WordVis. WordVis is an exceptionally agile, online synonym-dictionary explorer. It shows a customizable graph of interlinked words and meanings, which are arranged by real-time animation, and behave like repelling magnets that are also rearrangeable by the user. WordVis shows the merits of using JavaScript: the visualizer can be fully integrated inside a web page, and that enables communication with other HTML elements in the page. It demonstrates the latest developments in web technology, and its agility may serve as an inspiration for other web-applications. As an intuitive, online synonym-dictionary explorer, it helps to ‘find the right word’ while writing English texts. It offers an online, free and more powerful browsing experience than any existing word-network visualizer.

This paper proposed a Reduce Identical Event Transmission Algorithm (RIET). The algorithm can decide that which sensor nodes could send the event to sink node when sensor nodes sense a same even. Moreover, other nodes can save power because they didn’t send the same event. In our simulation, the RIET algorithm can enhance sensor nodes’ life time about 12.9 times and saving power consumption about 52.43 % than tradition algorithms.

Business process management and improvement are vital for enterprises in competitive environments. Understanding of a process is a pre-requisite and important step for improvement. Interaction between humans, computers, and business objects provide excellent opportunities for knowledge extraction. However, the specification of a framework is required for business process improvement, which extends from data collection, analytical methods, storage, and representation of knowledge. The process models conceived for information system development are not sufficient for post execution analysis and improvement. In this paper, we specify such a framework briefly and focus on providing representational support for business process improvement. The main objective is to improve the overall improvement process by providing enriched graphical process models. Furthermore, we use a case study to explain the proposed usage and extensions of an existing modeling language for business process improvement.

This paper proposes an automatic localization and boundary detection of retina images using basic filters to support ophthalmologists for detection and diagnoses eyes harmful diseases such as glaucoma and diabetic retinopathy accurately and diligently. The proposed system comprising three main phases including preprocessing, segmentation and detection phase. The preprocessing phase is used to enhance retinal image and to remove the noise of the retina image. The second phase is the segmentation for main parts of retinal image including optic disc, blood vessels, and fovea to extract their features. Optic disc is segmented using color intensity, the region of interest (ROI) is detected and morphological operations are applied to reduce search complexity. Also, fovea feature is extracted and the blood vessels tree is extracted from retinal image using line detection techniques. The third phase is the detection, in which identification and classifying whether the input image is left or right eye, to support ophthalmologists in identifying which eye is infected by the disease and to check it periodically. Basic image processing filters including average filter, median filter, spatial filter and morphological filter are used in all system phases. Moreover, a simple approach were used to detect left and right retinal fundus images. The proposed system is tested and evaluated using a subset of ophthalmologic images of the publically available DRIVE database.

The age of the population in all societies around the world is increasing. Elderly people prefer to maintain their independence, their autonomy and live at home as long as possible. This research take place in the field of the health care telemonitoring system which proposes software solutions to monitor elderly people in their own home. This study aims to make profit of the technological diversity of the several Decision Support System used to detect distress situation. We propose a multi-Agent approach to federate a collective decision process in which each Agent encapsulates a decision support system. This encapsulation enables the real time combination of the decisions.

There is a need for handling preferences in relational query languages that arises naturally in real-world applications dealing with possible choices generated by the current state of the world captured in the relational data model. To address this problem, we propose a fully declarative language for encoding preferences conditional on the current state of the world represented as a relation database instance. The language has constructs for various kinds of preferences, and we show how to interpret (sets of) its formulae; even sets of formulae that encode conflicting preferences. This leads to a flexible approach for specifying the most desirable choices of autonomous systems that act on behalf of their designers. Throughout the paper, we use an example of a control support system for a bank surveillance to motivate the need for our framework and to illustrate it.

In this paper we evaluate various approaches to a user profile modelling for news recommendation. We represent a user profile as a bag of real world entities, the user is interested in. News articles are thus recommended based on its contained concepts and not based on a text similarity. We propose several ways of such a user profile construction based on a user feedback. Different ways of a user feedback collection are compared. This paper addresses the problem of precise user modelling for information filtering.

Modern graphics cards take role of powerful computation hardware. This hardware becomes more popular due to purchasing costs and its availability. The advantages of Graphics Processor Unit (GPU) in parallel computation of Self-Organizing Network are described in this paper including a comparison with multi-threaded CPU. The parallelism on GPU is explained in a separated section. Mentioned section is divided into parts with respect to different forms of parallelism. The results of experiments at the end confirmed, that the utilization of GPU brings significant improvements in time of computation in case of large data sets.

Boolean factor analysis is one of the most efficient methods to reveal and to overcome informational redundancy of high-dimensional binary signals. In the present study, we introduce new Expectation-Maximization method which maximizes the likelihood of Boolean factor analysis solution. Using the so-called bars problem benchmark, we compare efficiencies of the proposed method with Dendritic Inhibition neural network.

In this paper, we present a generic interactive framework based on human cognition, where the system can learn continuously from the Internet and from its interaction with the users. To show the utilization of this framework, Iintelli, an agent based application for multiple text document summarization is developed and compared with the MEAD on the Cran Data Set. Mead is a natural language processing-based summarizer, which provides summary by extracting sentences from a cluster of related documents and Cran is a data set maintained by Information Retrieval Group at University of Glasgow. The human knowledge and experience are represented through fuzzy logic-based word-mesh and sentence-mesh, which can learn. Learning is performed using the competitive models, namely, Maxnet and Mexican Hat Models. As the result shows, the framework performs well as a multi-document summarizer. Though we have tested the framework for multi-document summarization, we believe that it can be extended to develop interactive applications for other domains and tasks.

This article introduces a hybrid scheme that combines the advantages of pulse coupled neural networks (PCNNs) and support vector machine, in conjunction with type-II fuzzy sets and wavelet to enhance the contrast of the original images and feature extraction. An application of MRI breast cancer imaging has been chosen and hybridization scheme have been applied to see their ability and accuracy to classify the breast cancer images into two outcomes: cancer or non-cancer. In order to enhance the contrast of the input image, identify the region of interest and detect the boundary of the breast pattern, a type-II fuzzy-based enhancement and PCNN-based segmentation were applied. Finally, wavelet-based features are extracted and normalized and a support vector machine classifier were employed to evaluate the ability of the lesion descriptors for discrimination of different regions of interest to determine whether they represent cancer or not. To evaluate the performance of presented approach, we present tests on different breast MRI images.

Very interesting source of information about scientific publishing in computer science is database DBLP. This database allows bibliographic information about main publications from conferences, journals and books in this area. In the article we deal with strength extraction between authors based on their association. The research presented in this article is partly motivated by work of Mori et al. From this paper we have used the approach for extraction of initial metadata, and we have inspired how to take advantage from Jaccard coefficient principals for description of the strength of associations between authors. Method is usable for development of synthetic coauthors network, where as input is used the set of words, which will describe the network (the authors used these words in publication titles).

This paper is focused on the detection of communities in social networks. We propose and describe a novel method for detecting local communities. We have used this method in an experiment on student social networks in order to prove our hypothesis about the nature of student communities. The results of the experiment rationalized our hypothesis and confirmed the effectiveness of the described method of local community detection.