Dimitrios Tsonos

Multimodal Accessibility of Documents

Traditionally a “document” was considered as a textual record. Schamber (1996) defined document as a unit “consisting of dynamic, flexible, nonlinear content, represented as a set of linked information items, stored in one or more physical media or networked sited; created and used by one or more individuals in the facilitation of some process or project”. Buckland (1997) tries to answer the question “what is a document?” through a discussion on how far the meaning of “document” can be pushed and which are the limits of “documentation”. The evolution of Computer Science and Information Technology created a new perspective for the word “document”. The concept “electronic documents” can be differentiated from the “printed documents” having specific characteristics (Schamber, 1996): easily manipulabable, internally and externally linkable, readily transformable, inherently searchable, instantly transportable and infinitely replicable. The term document used in this chapter refers to all kind of printed or electronic documentation the content of which is most text, like newspapers, books, journals, magazines, educational material, letters, brochures, leaflets, etc. A document contains elements that arrange content in the page or even in the document itself. For example, the title of a chapter can be recognized by placing it on the top of the page and in larger font size than the body of text. Also, page indexing at the end of a document links the reader to specific part of the document. The document functionality can be distinguished into browsing, searching, navigation and reading (Doermann et al., 1998). Additional concepts related to document’s functionality are legibility, readability and aesthetics. Legibility is a measure of easiness to distinguish one letter from another in a particular typeface and readability is a gauge of how easily words, phrases and blocks of copy can be read. These two measures were introduced as a comparison between printed and electronic documents on a computer screen (Mills & Weldon, 1987). Readability and legibility are closely related to typographic elements, typeface design and font/background color combinations (Hill & Scharff, 1997; Richard & Patrick, 2004; Eglin & Bres, 2003). Readability is more related to the overall and layout structure of a document (Holmqvist & Wartenberg, 2005; Holmberg, 2004; Kupper, 1989; Wartenberg & Holmqvist, 2005; Axner et al., 1997). Aesthetics of a document play a significant role during the reading process as well as reader’s preferences (Porat et al., 2007; Harrington et al., 2004; Laarni, 2003). “Dynamic” concepts like navigation and browsing are related to the interaction process with the reader. Navigation in electronic documents is a set of instructions to create an appropriate

Auditory accessibility of metadata in books: a design for all approach

There are two issues that are challenging in the life-cycle of Digital Talking Books (DTB): the automatic labeling of text formatting meta-data in documents and the multimodal representation of the text formatting semantics. We propose an augmented design-for-all approach for both the production and the reading processes of DAISY compliant DTBs. This approach incorporates a methodology for the real-time extraction and the semantic labeling of text formatting meta-data. Furthermore, it includes a unified approach for the multimodal rendering of text formatting, structure and layout meta-data by utilizing a Document-to-Audio platform to render the acoustic modality.

Towards modeling of Readers’ Emotional State response for the automated annotation of documents

This work presents an experimental study towards modeling the readerspsila emotional state as a response to font and typesetting elements of documents presented on a LCD display. Any content and/or domain dependent information was excluded from the document that was tested. An automated computer-based experimental procedure has been followed based on the paper-and-pencil self assessment Manikin test. The typographic elements: font color, size, type, background color and the typesetting elements: bold, italics, bold-italics, along with their combinations are studied. The results indicate that the combination of text and background color has the same impact across languages; the font size has a typical behavior. Readerspsila emotional state, induced by the typesetting elements and the font type, probably depends on the current as well as on the previously displayed stimuli. A cognitive-based XML architecture for real-time extraction of readerspsila emotional state relatively to documentspsila typographic elements is also presented. The results of this paper can be considered when transforming emotionally annotated documents into acoustic modality.

A Methodology for the Extraction of Reader's Emotional State Triggered from Text Typography

Writing is employed by humans in order to communicate, exchange ideas or store facts and descriptions. A well known Latin phrase is “verba volant, scripta manent” i.e. “spoken words fly away, written words remain”. Humans in prehistoric years, used figurations to describe several events and express their fear or their admiration, the ancient Egyptians used papyrus to write down their ideas and the Chinese, in 11th Century, went beyond handwriting using typography with moveable type to create multiple copies of their documents. Through centuries typography has evolved and nowadays is one of the major manners to exchange ideas and information. Using computers, typography and the way we write has changed radically. We have a plethora of software tools, e.g. for writing a plain text (like NotePad in MS Windows), word processors (e.g. Open Office) for text editing as well as for more elegant visual appearance of the documents and professional tools for the creation of posters, advertisements e.t.c. (e.g. Adobe’s Tools and Editors). Utilising these tools, computer users have the ability not only to write a plain text, but also to format the text and arrange it in the page. In newspapers, for example, the page designer distinguishes the title from the body of text at the top of the page or the article, with larger font size. Also, when an editor wishes to emphasize a specific word or phrase he uses bold or italics typesetting. A writer can convey a message, a felling or an idea not only by the meaning of the content but also by the way the text is visually presented to the reader. Page layout affects the way a newspaper is read (Holmqvist & Wartenberg, 2005; Holmberg, 2004; Kupper, 1989; Wartenberg & Holmqvist, 2005). The use of the WWW and the web page creation and design, introduced a new perception for the meaning of documents and publishing. In web pages, the text and background color combinations have impact on the readability and aesthetics (Porat et al., 2007; Richard & Patrick, 2004; Hill & Scharff, 1997) and a well designed graphical web document can be reader friendly (Borchers et al., 1996). Humans express their emotions in every personal or social occasion. Everything they do or make is followed by or follows an emotional expression. For example, some people are afraid of being among many people. So they act accordingly by avoiding congestion. If they cannot, they are feeling anxious and nervous. Another example of the role of the emotions in

Design and Development Methodology for the Emotional State Estimation of Verbs

The use of words and particularly the verbs in Human-Human Interaction reveals significant aspects of both human’s social and mental state. This work presents a novel methodology towards the emotional assessment of verbs by users. Essentially we would like to study whether the emotions that user experience are comparable with the corresponding results obtained through a mixture of natural language and statistical classifiers in SentiWordNet. Following the paper and pencil guidelines of the International Affective Picture System (IAPS) we have developed a web-based unsupervised version of the Self Assessment Manikin (SAM) test, designed for the emotional assessment of verbs in English and Greek language. Thirty five men and seventeen women participated in an internet survey version of the experiment. In the first part of the process, the participants had to assess their induced emotional state while reading a verb (totally 75 Greek verbs), on 5-point scales of “Pleasure”, “Arousal” and “Dominance”. The results comprise coherence and consistency. As a rule, all verbs obtained low to mid range scores on Arousal and Dominance axis and only on the Pleasure dimension scores are close to the edge.

Emotional Prosodic Model Evaluation for Greek Expressive Text-to-Speech Synthesis

In this study we introduce a novel experimental approach towards the evaluation of emotional prosodic models in Expressive Speech Synthesis. It is based on the dimensional emotion expressivity and adopts the Self-Assessment Manikin Test. We applied this experimental approach to evaluate an emotional prosodic model for Greek expressive Text-to-Speech synthesis. We used two pseudo-sentences for each of the Greek and English HMM-based synthetic voices, implemented in the MARY TtS platform. Fifteen native Greek participants were asked to assess eleven emotional states for each sentence. The results show that the “Arousal” dimension is perceived as intended, followed by the “Pleasure” and “Dominance” dimensions’ ratings. These preliminary findings are consistent with the results in previous studies.

Language Resources and Evaluation for the Support of the Greek Language in the MARY Text-to-Speech

The paper outlines the process of creating a new voice in the MARY Text-to-Speech Platform, evaluating and proposing extensions on the existing tools and methodology. It particularly focuses on the development of the phoneme set, the Grapheme to Phone (GtP) conversion module and the subsequent process for generating a corpus for building the new voice. The work presented in this paper was carried out as part of the process for the support of the Greek Language in the MARY TtS system, however the outlined methodology should be applicable for other languages as well.

User perception knowledge for socially-aware web document accessibility

Social Media provide a vast amount of information identifying stories, events, entities that play the crucial role of shaping the community in an everyday heavy user involvement. This work involves the study of social media information in terms of type (multimodal: text, video, sound, picture) and role players (agents, users, opinion leaders) and the potential of using that information for the design of accessible, usable preservation strategies. The challenge was to analyze the social web and present ways of preserving the web documents with social content in such way as to make them accessible for the future. The web documents should preserve accessible data and stored in such way as to enable intelligent retrieval.

Usability design and testing of an interface for search and retrieval of social web data

The vast amount of data on the web has been extensively harvested for many years for the purpose of digital archiving. In the recent years, however, the social networks contain the sources of most of the debating between the people. Recent approaches include social web information to the archived content for various reasons. This work reports on the usability design and evaluation of a search and retrieval user interface that was designed to retrieve web objects along with semantic information analyzed for the social web. The main task of the interface was to combine the social information with the standard archived content in meaningful and usable ways.

On the Identification and Annotation of Emotional Properties of Verbs

Adequate and reliable lexical resources are essential for effective sentiment analysis and opinion mining. This paper proposes a methodology for the emotional assessment and annotation of words. The process is based on the Self Assessment Manikin test, and is coupled with two psychometric measurements for identifying possible bias due to the annotator’s psychological condition and personality: the EPQ scale and the SCL-90-R scale. A web based tool was developed to support the process. The methodology was validated through a pilot study in which 10 participants were asked to assess the emotional state elicited by each of 75 verbs that were used as stimuli. Results are compared with SentiWordNet’s emotional scoring on respective verbs, and primarily show logical continuity and consistency.