Catálogo de publicaciones - libros

Web accessibility has become one of the most popular research targets. Web accessibility means the ability to be accessed by all kinds of people or devices. One problem that some Web pages have rather poor information structures at the HTML source code level, even though the pages are well structured visually. In order to transfer the visual structure to the HTML level structure, we need to identify the graphic design features that influence human understanding. We examined whether several people interpreted a displayed structure in the same way or not. Four subjects participated in trials; they attempted to identify the headers, major items forming the structure of the top page, of the web sites of 7 local governments in Japan. 80 % of the headers were selected by all subjects so most subjects shared the same understanding of what constituted a header.

A test suite is a set of tools and data to ensure quality of a software application such as a dynamic web site. Reliability of page by page web accessibility tests is difficult to achieve as this requires defining appropriate partitions of the set of all web pages. We propose a partitioning method based on additional information about reading behaviour by users especially for measuring readability and ease of navigation. An upper limit for the path length and hence size of the partition is developed. We present a test case editor to describe user group specific testing scenarios and additional types of tests.

This paper reports on the results from a series of psychoacoustic experiments in the field of the auditory representation of texts via synthetic speech which comprise similar acoustic patterns so called “paronyms”. The errors which occur when listening to paronyms are classified as errors of phonological type. Thirty blind and thirty sighted students participated in psychoacoustic experiments. The results from the experiments depicted the types of the subjects’ errors and addressed comparisons between the performances of blind and sighted students on their auditory distinctions towards the chosen scripts (paronym words and sentences with paronyms). The discussion considered the practical implications of the findings such as issues regarding education as well as the development of suitable design of acoustic rendition of texts in favor of better perception and comprehension.

Access to spoken information presents challenges in educational, work, and public settings for many individuals. Speech Recognition technology offers a potential strategy for increasing access in various settings, including the university lecture environment. An international research team has successfully experimented with providing speech recognition generated lecture notes, available online as multimedia enhanced transcripts. Additionally, a specially designed search and indexing tool was introduced to enhance the user experience. An exploratory study was conducted to gauge student reactions, identify usage patterns, and assess ease of use.

This study looked at continuous automated speech recognition (ASR) to an audience in a university lecture theatre and ran an evaluation based on a previous experiment by Ryba, McIvor, Shakir and Paez, which found that non-native speakers of English were much more favorable towards the use of ASR in class than native speakers. Our evaluation was done on a class of 29 students composed entirely of non-native speakers/ A strong indicator of the level of engagement with the technology was the linguistic ability of the user – the weaker the student’s English, the more he or she tended to look at the textual output, the greater distraction experienced through poor recognition and also the greater impatience felt with slow recognition.There also seemed to be cultural differences – the Chinese students appeared to look at the textual output much more than Indian students. We conclude that the 2 axes around which successful classroom speech recognition occurs are those of accuracy and unobtrusiveness. The more accurate and unobtrusive the technology, the more successful will be the automatically transcribed lecture.

Learners with disabilities experience many difficulties when reading. This study aimed to develop a TriAccess computer system to assist learners with disabilities to read effectively. TriAccess provide physical, sensory, and cognitive supports to learners. The system also offers professionals a convenient and web-based interface on material development and system management. In addition, the study conducted a usability evaluation on the TriAccess. Thirty participants involved in the evaluation process. The results have shown that the TriAccess system is valuable not only to learners, but also to instructors, and material developers.

This paper described a system which could measure cursor movement and control site motion simultaneously. In addition to evaluate a client’s pointing and selecting proficiency, CAT Motor system, developed on the basis of the previous CAT system, was also used to detect the motion of his control site. This feature could provide clinical professionals more detailed information when comparing the difference of performance among devices. The CAT Motor system consisted of four basic types of mouse operating skills, including clicking, cursor moving, cursor moving and clicking, and dragging. The CAT Motor recorded the location and the response of the pointing device and captured the picture of the control site synchronally. Then, the system visualized the results of cursor moving and control site motion on the screen. The CAT Motor system not only provided the professionals with an objective mean of evaluating specific difficulties of mouse operating difficulties, but also allowed them to evaluate the effort of the motor control training. Besides, this paper had some suggestions for the research in the future.

In this paper, we outline a cognitive architecture for communicators, called the BIC model. The model consist of three main components. First, a (B)iological component, in which the genetic or built-in capacities of the communicator are specified. Second, an (I)nteraction Engine which uses neo-Gricean mutual simulation to attribute communicative intentions to signals, and to create signals to convey communicative intentions to the I-system of other agents. The third component of the BIC model is a content addressable database of (C)onventions which is used to store form/meaning mappings that have been successfully computed by the I-system. These stored form/meaning mappings are indexed by types of communicative context, so they can be retrieved by the I-system to save computational resources. The model can be used both as a computational architecture for a communication module in an artificial agent and as a conceptual model of the human communicator.

This paper presents the results of a study comparing the use of auditory icons, earcons and speech in an audio only interface for a digital talking book player. The different techniques were evaluated according to the identification errors made, and subjective measures of understandability, intrusiveness and pleasurability. Results suggest the use of auditory icons combined with speech whenever necessary, in detriment to the use of earcons, for applications sharing the characteristics of digital talking book players.

A speech interface sometimes provides easy access to WWW information since it makes browser potentially more friendly and powerful. This study first made a comparison between a voice and a keyboard input task. It then compared the voice interface with the keyboard, mouse, and combination of keyboard and mouse with three WWW browsing tasks. The experimental results showed that the voice interface performed the input task faster than the keyboard. The voice interface reached an input accuracy of 92% after sufficient learning. In the WWW browsing tasks, the voice performed poorly in terms of completion time, compared to the keyboard, mouse, or both. Nevertheless, the voice interface can be used to complete the three assigned tasks without problems. The study showed that voice interaction can be used as a universal web interface, especially for those who are unable to use the mouse and keyboard due to disabilities.