The ultimate aim of research into computer vision is designing a system which interprets its surrounding environment in a similar way the human can do effortlessly. However, the state of technology is far from achieving such a goal. In this thesis different components of a computer vision system that are designed for the task of interpreting man-made scenes, in particular images of buildings, are described. The flow of information in the proposed system is bottom-up i.e., the image is first segmented into its meaningful components and subsequently the regions are labelled using a contextual classifier. Starting from simple observations concerning the human vision system and the gestalt laws of human perception, like the law of “good (simple) shape” and “perceptual grouping”, a blob detector is developed, that identifies components in a 2D image. These components are convex regions of interest, with interest being defined as significant gradient magnitude content. An eye tracking experiment is conducted, which shows that the regions identified by the blob detector, correlate significantly with the regions which drive the attention of viewers. Having identified these blobs, it is postulated that a blob represents an object, linguistically identified with its own semantic name. In other words, a blob may contain a window a door or a chimney in a building. These regions are used to identify and segment higher order structures in a building, like facade, window array and also environmental regions like sky and ground. Because of inconsistency in the unary features of buildings, a contextual learning algorithm is used to classify the segmented regions. A model which learns spatial and topological relationships between different objects from a set of hand-labelled data, is used. This model utilises this information in a MRF to achieve consistent labellings of new scenes.

No description available.

This thesis investigates human factors issues in the design and development of effective human-robot interfaces for emerging applications of teleoperated, cooperative mobile robot systems in situations such as urban search and rescue. Traditional methods of designing human-robot interaction interfaces have failed to produce effective results as witnessed in the post September 11 search operations. The thesis adopts a user-centric approach based on the human factors of situation awarenes~, telepresence and workload to support the hum'an operator because this is widely accepted as the best way of realising increased levels of collaboration between humans and robotic systems, working as a partnership to perform a complex task. The measurement of these human factors has not been explored within the robotic community in any significant way. The measurement of these subjective and complex issues is addressed in this thesis by looking to the flight traffic control domain where researchers have developed many methods of determining how to quan- .- - .. - ._-------------_._---------------_ ..-----_.._--- tify the quality of situation awareness, the level of workload and the level of telepresence of the people in the aircrafts and on the ground. Based on these methods, the research proposes five new methods (ASAGAT, QASAGAT, CARS, PASA, SPASA) for measuring situation awareness, three methods (WSPQ, SUSPQ, SPATP) for measuring telepresence and three methods (NASA-TLX, MCRS, FSWAT) for measuririg workload. A comprehensive comparison between them has shown that QASAGAT and SPASA are the most reliable and accurate for measuring situation awareness, SPATP for measuring telepresence and FSWAT for measuring workload. For the measurement of performance a new method has been developed, which is felt to be more objective for the urban search and rescue scenario than the metrics used in the RoboCup Rescue competition. Simulation studies involved extensive investigations to determine the various software tools and platforms that are available for realising robot urban search and rescue scenarios. The software of Player-Gazebo was selected as the best solution. A graphical user interface· comprising vision, laser data, map, robot locations, etc. was developed and assessed with the proposed measurement methods under the simulated robot system and search scenarios. The test subjects comprised specialist end users as well as general non-end users. . An in-between groups analysis showed that the individual characteristics of each group may have some effect on the experimental variables, however, this effect is very minimal and the l?ain influence factors are the interaction interfaces and the human factors investigated here. Moreover, the results indicated that there is no significant benefit when using professional urban search and rescue end users. Correlation analysis on the data has shown that situation awareness and telepresence have a positive effect on performance, while workload negatively affects performance. It was also found that there is a positive correlation between situation awareness and telepresence, while workload has a negative effect on both. These results validate the assumptions made. __________Amultiple.linearregressionmodel has_been.developed _to further _l.lnderstand.the _ individual contributions of each human factor in the overall performance achieved. The limited prediction capabilities of th'-e linear model suggested anon-linear relationship. For this reason,'- a non-linear model using an artificial neural network trained with the backpropagation algorithm has also been developed. The resulting neural network is able to predict the response variables more precisely and is shown to be able to generalise well to unseen cases. A physical mobile teleoperated urban search and rescue robot system has also been developed for realising future real world trials.

Existing dialogue systems can be categorized into three main types based on their styles of questioning, namely the Finite-State, Frame-based and Free-Form strategies. Finite-State presents a list of valid options to the user in the system questions and expects the user to reply with one of the options. Slot-filling Frame-based provides a flexible approach for the system to capture any useful data that are related to the current context from the user's reply. Open-ended Free-Form is based on the questioning style used in agent-based systems, which allows the system to analyse the user's input with regard to the whole domain of the system rather than only the current context. Each of the three dialogue strategies has its own benefits in different circumstances. Therefore it is desirable to have a dialogue system that can use all three strategies in different circumstances. With such a dialogue system it is possible to use a flexible strategy such as the Frame or Free-Form in normal or optimal situations, but use the more step-by-step Finite-State strategy in situations where the system demands more control on the dialogue. The algorithm for determining the selection and transition of dialogue strategies is based on what are regarded as reasonable human behaviours. The algorithm aims to conduct conversations that are natural, efficient and able to adapt to errors of misunderstanding or misrecognition according to the dialogue history of the dialogue session and the individual user. A software application that uses the algorithm has been developed and tested for user satisfaction by a number of testers with various level of experience in using dialogue systems. Test results show that the testers had better dialogue efficiency and user satisfaction with the system that uses the multi-strategy dialogue manager and that the majority of the testers prefer to use such a system. Supplied by The British Library - 'The world's knowledge'

This thesis employs a user-centered approach to design and develop an extensible, adaptive system, the ACTIVE system to improve Web graphics accessibility for visually impaired people. It is capable of adapting to its context of use such as graphics type, asssistive technologies and user profiles in order to choose a suitable graphical multimodal application for the user. Additionally, the system consists of a user model which contains information about a user in terms of their background, experience levels, and preferences. By performing a series of experiments with visually impaired people, it reveals that people with similar profiles and experience levels prefer certain exploration conditions. Consequently, a list of adaptation rules have been derived and applied in the system. By using a feature-based approach, the ACTIVE system learns about the users from their previous interaction with applications and presents to them with their most preferable and appropriate interface. The ACTIVE system was designed in accordance with usability and accessibility guidelines. The system was evaluated with visually impaired people and the results reveal that it has improved the overall experience and satisfaction of people with sight loss when accessing graphics non-visually. Furthermore, the adaptation accuracy level of the ACTIVE system increases with the degree of system use, where 96.67% of accuracy was achieved in the experiments. This thesis shows the possibilities in developing a coherent, adaptive system by integrating various variables such as graphics types, assistive technologies and multimodalities. It also demonstrates that adaptation can bring benefits to people with visual impairments in enhancing their graphics accessibility.

No description available.

No description available.

No description available.

No description available.

This work explores the use of Interactive Visualisations to solve Exam Timetabling problems by making use of new graphical representations and interactions. Exam Timetabling is a. well researched topic in optimisation, however existing automated solutions are limited in terms of interactivity and generality. In cognitive science relevant literature shows how different representations can affect the outcome of problem solving tasks yet existing timetabling systems and research focus almost exclusively on automated approaches. not making use of available human abilities. These shortcomings of current approaches are addressed. raising the questions: Can effective representations be used to completely solve Exam Timetabling problems without use of computer automation and can such representations be effectively integrated with existing approaches? Previous research suggests that closely mapped expressive representations can be effective problem solving tools. In this research representational design methodologies were applied to develop two systems capable ofsolving complete problems. rather thanjust improving existing solutions. VAST (Visual Analysis and Scheduling for Tl1Detabling) provides interactive visualisations to solve Exam Timetabling Problems and introduces novel Cluster Group visualisations to simplify the problem faced by users. KNIGHT (Knowledge Interactive Guided Heuristic Timetabling) extends these interactive visualisations to include human in the loop interactivity integrated with existing automated heuristics. The application of mobilities to the domain of Exam Timeta~ling allow heuristic searches to be guided toward new solutions, expanding potential search neighbourhoods whilst excluding solutions that are perceived to be problematic. An evaluation of KNIGHT is presented which discusses this approach. The evaluation of VAST provides positive evidence for the use of interactive visualisation as an integral part of the problem solving process and our analysis shows the discovery of different user strategies that make use of these new interactions and representations. Whilst the use of exam clusters appears beneficial no significant benefits were found from allowing initial exam assignments.