Facebook Messenger, ChatSecure, Line, QQ and WhatsApp represent examples of Mobile Messaging Applications (MMAs) that collectively facilitate instant and global communications, in which billions of audio, image, text and video messages are exchanged on a daily basis. Some of these messages are digital assets (messages that have a significant functional, sentimental or monetary value) that may be exposed to confidentiality or exploitation threats, wehn published in MMAs that are subject to such security concerns. Additionally, some MMAs impose message format constraints that prevent users from publishing sudio, image and video digital assets in an encrypted form, thus hampering their attempts to mitigate such concerns. In this thesis, two research contributions are presented, that are designed to establish confidentiality over digital assets that are messaged via MMAs. The first contribution, Dual Carrier Steganographic Messaging (DCSM) protocols, pairs two independent and non-colluding MMAs as communication channels, in which a digital asset is steganographically messaged over one MMA and its stego-key (size, format, access information) over the other. The second contribution, extends a probabilistic Hash-based Least Significant Bit (HLSB) steganography technique through the introduction of hash function arrays, and can be applied to JPG images in the spatial domain. This technique features as an integral component in DCSN schemes. Both research contributions are implemented and evaluated via a prototype application, Conceal[sub]2. JPG HLSB embedding is evaluated in a number of configurations that are applied to cover-images with diverse pixel luminance profiles. Subsequent performance results are measured in terms of embedding overheeads, carrier-image distortion ranges, thresholds and embedding scope. Results conclude that JPG HLSB embedding is most efficient when utilising images with wide variations in pixel luminance, and that a configuration utilising an array of 16 hash functions, delivers the best performance of the array sizes tested. Empirical evaluations of 54 Android MMAs conclude that their use as communication channels in DCSM schemes, is feasible where MMAs maintain image integrity and that the successful act of steganographic messaging over viable MMAs without subsequent corruption of carrier-images, indicates that carrier-images are detected but only observed, or that they are not monitored, or that MMA systems are ineffective at detecting the carrier-images utilised in this research.

We present in this thesis the design and development of a new mobile diabetes management system for social behavioural change and management tailored for Saudi diabetic patients (SANAD - Saudi Arabia Networking for Aiding Diabetes). The hey goals of SANAD are to close the diabetes management loop by providing remote monitoring for diabetic patients, a further therapeutic channel to the patient, an opportunity to increase diabetic patients' health awareness, and feedback to help diabetic patients maintain a regular blood glucose level. The key system components consist of: (i) a smart mobile diabetes management module (MDM-M) used for collecting blodd glucose data; (ii) a social networking module (SN-M), acting as an enhancement module for the MDM-M, the key function of which is currently focused on education purposes; and (iii) a cognitive behavioral therapy module (CBT-M),a cting as a supplementary module to MDM-M. This module was designed on the smart mobile platform and used only by patients who require CBT therapeutic intervention. A usability study for the SANAD system is also presented in this thesis to validate the acceptability of using mobile technologies amongst diabetic patient in KSA and Gulf region. The preliminary results of the study indicated general acceptance of the patients in using the system with higher usability rating in type 2 diabetic patients. In general, the study concluded that the concept of SANAD system is considered acceptable tool in particularly with Type 2 diabetes patients. A clinical evaluation study of SANAD system is also conducted in this thesis to evaluate the clinical effectiveness of the system. The primary aim of this study was to evaluate the effect of SANAD system on: 1) improve glycaemic control; 2) improve health awareness; and 3) enhance self-efficacy. Secondary aims qualitatively evaluated the content of communication in SANAD system. The key preliminary results of this study provided in evidence that SANAD has a positive impact on promoting knowledge of diabetes in individuals living with type 2 diabetes, and reflects the generally positive outcomes of reducing glycated hemoglobin control (HbAlc (%)), and increasing self-efficacy.

Intelligence is the power which makes the owner capable of making a decision defined by reasoning. When traditional solutions and approaches, such as First Principal Modelling or Statistical Modelling, are not feasible or able to effectively address complex real- world problems, then Computational Intelligence with some nature-inspired computational techniques and methodologies is employed. For transferring data between two non-directly connected devices when some other devices are in-between, a set of rules are used by routers which are devices between sender and receiver, to determine the most appropriate paths into which routers should forward data toward the intended destination. This set of rules is called routing protocol. Researchers use some computational itelligence techniques to design network routing protocols. Wireless Sensor Networks (WSNs) play an important role in today's data communication systems and researchers are expected to proliferate in the field of wireless communication in the near future. The deployment of wireless sensor networks offer several advantages in comparison to traditional infrastructure-based networks, such as fully distributed mobile operation, the easy discovery of joining wireless devices, and instant and low cost network setup. Designing an effective routing protocol is one of the main challenges in the ad-hoc networking paradigm and the utilisation of an adequate link-cost metric is essential. WSN researchers address issues such as low throughput and high latency in wireless sensor data communication. Routing Protocols in WSNs play a key role in data communication and the main parameter in all routing protocols is data communications link-cost. This research delivers two surveys on existing routing protocols and link-quality metrocs for wireless sensor networks. Most of the routing protocols in this area are considered in different groups. The majority of link-quality metrics in WSNs are studied in different categories. Link-quality and traffic-aware metrics account for most of the metrics, as well as metrics in multi-channel networks and cognitive radio systems, which are also considered in detail. Metrics are reviewed in detail in terms of their performance; summary and comparison tables of link-quality metrics are provided to enable better comparison and show a brief overview of their appearance to get a clearer picture. Routing-metrics are important is determining paths and maintaining quality of service in routing protocols. The most efficient metrics need to send packets to maintain link-quality measurement by using the Radi Frequency (RF) module. In this study, a set of statistical analyses is done on some link-quality metrics to select the best metric for energy-aware scenarios. Two prominent link-quality metrics; Received Signal Strength Indication (RSSI) and Link-Quality Indication (LQI), are described in detail. The symmetry of RSSI and LQI in two directions is studied, and relations with the Expected Transmission Count (ETX), RSSI, and LQI as link-quality metrics are analysed. The evaluation in this research is based on a series of WSN test-beds in real scenarios. Due to implementation of routing protocols in limited power supply devices in WSNs, one novel link-quality metric and also some routing protocols for wireless sensor networks are proposed in this research to obtain better performance in different scenarios. Rainbow Collection Tree Protocol (RCTP) is presented and evaluated as an enhanced version of Collection Tree Protocol (CTP). It uses the Trickle algorithm to optimise overhead cost and the algorithm also makes RCTP quickly adaptable to changes in topology. The Rainbow mechanism is used in RCTP to detect and route around connectivity nodes and avoid routes through dead-end paths. Energy-efficient Rainbow Collection Tree Routing Protocol (ERCRP) is presented and evaluated as a novel, real-time, position-based and energy-efficient routing protocol in this research. ERCRP is a lightweight protocol that reduced the number of nodes which receive the RF signal using a novel Parent Forwarding Region (PFR) algorithm. ERCRP as a Geographical Routing Protocol (GRP) reduced the number of forwarding nodes and thus decreases traffic and packet collision in the network. WSNs are used in three-dimension (3D) scenarios such as sea or land surfaces with different levels of height. Three-Dimension Position-Based Adaptive Real-Time Routing Protocol (3DPBARP) is presented and evaluated as a novel, real-time, position-based and energy-efficient routing protocol for WSNs in this research. 3DPBARP is a lightweight protocol that reduces the number of nodes which received the RF signal using a novel PFR algorithm. 3DPBARP as a GRP decreases the number of nodes which participate in packet forwarding and thus shrink the traffic and collision in the network.

While collaborative learning is universally recognised as a process of considerable pedagogical value, and the enhancement of standard application interfaces to promote collaboration has formed the subject of previous studies, the literature addressing the degree to which Multi-touch technology is used remains limited. Multi-touch, multi-mouse tools are a novel category of groupware, which can enhance learning through collaborative use of those tools. The research aim is to develop a framework and a software tool known as the Learning to Program through a Computer-Aided Collaboration (LPCAC) tool. The LPCAC is a Multi-touch software tool that enhances the learning of computer programming in university through collaborative use of this tool. Subsidary objectives are pursued in the form of research questions, with most comparatively investigating and evaluating the use of the Multi-touch tool compared to the standalone PC-based approach. University students were selected to take part in a study to measure their performance improvements using the Multi-touch tools in a collaborative environment. A two-part process was followed to analyse the difference between student collaborations using Multi-touch with multi-mouse devices cmpared to standalone PC-based system with a single mouse device. Data collection involved video recording and a questionnaire analysis. Data analysis followed an eight-step process to evaluate the process of collaborative programming, which includes an analysis of collaboration patterns, time on task, equity of participation and design quality. While the Multi-touch tool is limited in a number of ways, its multiple points of contact and the collaborative activity promote a more favourable software design process when compared to the standalone PC-based condition. Although further research is needed, this study has found that Multi-touch technology represents a promising prospect for application in teaching computer programming in educational establishments, as well as improving the software development process in the wider collaborative software development.

Technology has transformed many aspects of business and market activities. Internet is one of the most important technologies, which have created a global digital economy with new opportunities. One of the most noticeable changes of business is e-commerce. E-commerce enables business to sell products and services to consumers on global basis. A prominent role of electronic commerce is that it assists firms to compete, have access to new markets and extend the geographic reach of their operations. An adoption of appropriate technology can lead the company to greater business competency, improve its business performance, and ensure it retains its competitive advantages. Despite this awareness, many SME in the developing countries in the Middle East have yet to fully embrace this adoption of ecommerce in the running of their business. The problem in this research is to identify and bring together in one framework the appropriate issues, variables, components and concepts that need to be addressed to encourage, motivate and enable SME's in Middle Eastern countries to adopt electronic commerce technology in the running of their businesses. The existing innovation adoption models are inadequate in appropriately explaining which factors are involved in the adoption decision, and are particularly insufficient with respect to small and medium sized enterprises (SMEs) and further, less research is available in the context of Middle Eastern countries. The aim of this study is to develop a comprehensive research framework called Electronic Commerce Technology Acceptance (ECTA) framework, which can be utilised for discovering the factors affecting the adoption of e-commerce innovation and to apply this framework for empirically testing the adoption of e-commerce application by SMEs in Middle Eastern Countries. The ECTA framework contains fourteen potential determinant factors covering four phases: social stimulus, cognitive response, affective response, and behavioural response. This empirical study was conducted through survey research and the sample was drawn by means of systematic sampling technique. The empirical data were collected by using self-administrated questionnaires and semi-structured interviews. The data analysis was based on 400 Jordanian SMEs; also data analysis was based on multivariate statistical techniques encompassing multiple linear regression, simple linear regression, one-way ANOVA, and stepwise regression. Miles and Huberman approach and Leximancer software were used for the qualitative data analysis. The findings of the study reveal significant insight into understanding the adoption of electronic commerce by SMEs.This research has developed, presented and evaluated the ECTA framework. This framework has been shown to be very viable and useful. The framework can he used as a tool to enable Technology innovation and adoption scholars to encourage and advising SMEs to adopt appropriate ecommerce technology to improve their business performance. Suggestions are made to improve the framework with future work to apply this framework in more SMEs. A strategy is suggested for policy makers to make use of these findings to improve the adoption of ecommerce in Jordon. The findings from this research show that the ECTA framework is beneficial to both governmental and private sectors who intend to accelerate the adoption rate of electronic commerce implementations and their relevant components among SMEs in Jordan.

This dissertation contributes to the state of the art in the field of pattern recognition and machine learning by advancing a family of nonlinear dimensionality reduction methods. We start with the automatisation of spectral dimensionality reduction approaches in order to facilitate the usage of these techniques by scientists in various domains wherever there is a need to explore large volumes of multivariate data. Then, we focus on the crucial and open problem of modelling the intrinsic structure of multidimensional time series. Solutions to this outstanding scientific challenge would advance various branches of science from meteorology, biology, engineering to computer vision, wherever time is a key asset of high dimensional data. We introduce two different approaches to this complex problem, which are both derived from the proposed concept of introducing spatio-temporal constraints between time series. The first algorithm allows for an efficient deterministic parameterisation of multidimensional time series spaces, even in the presence of data variations, whereas the second one approximates an underlying distribution of such spaces in a generative manner. We evaluate our original contributions in the area of visual human motion analysis, especially in two major computer vision tasks, i. e. human body pose estimation and human action recognition from video. In particular, we propose two variants of temporally constrained human motion descriptors, which become a foundation of view independent action recognition frameworks, and demonstrate excellent robustness against style, view and speed variability in recognition of different kinds of motions. Performance analysis confirms the strength and potential of our contributions, which may benefit many domains beyond computer vision.

The word processor or some form of editor-based application has become an integral tool for the many people who rely on computers on a daily basis. As such it has a wide and varied user base and must cater for a very diverse user group. Due to the heavy reliance on the word processor it is essential that it delivers pleasurable and efficient interaction to its users. Since its inception, the word processor has displayed the ability to evolve to continually exploit the increasing capabilities of technology. This study focused on furthering the improvement of the word processor usability for a subset of South African word processor users. Specifically, it concentrated on the impact of graphics, text and language on the usability of a word processor. Graphics were incorporated into the interface by means of inclusion of the icons currently found in the Microsoft Office package, which have been accepted as the industry standard, and the development of an alternative set of icons whose usability could be compared to that of the standard icons. Text was included in the interfaces in the form of menus and tooltips as well as text buttons which replaced the afore-mentioned pictorial icons and contained no graphical depiction of the associated function. The impact of language on the usability of a word processor was viewed strictly in terms of bilingual users and was achieved through translation of the text buttons, menus and tooltips into the predominant languages of the area. Comparative user testing was conducted through implementation of a scaled-down word processor application which could accommodate interchangeable interfaces and easy administration of preset tasks. Representative users were then required to complete a series of tasks on their respective pre-assigned interface, which conformed to one of the following general interface configurations: a. An interface using either set of pictorial icons and excluding both menus and tooltips, thus containing no language component. b. An interface in their first language, achieved through use of the text buttons, menus and/or tooltips. c. An English interface, where English was not their first language. A set of usability measures was identified which allowed for the effectiveness, efficiency and satisfaction of the users to be compared between the different user interface configurations. These measurements were: a. the score achieved for the test, based on a built-in difficulty index assigned to each task; b. the satisfaction experienced during interaction with the application; and c. for each task, the: i.time, ii.number of actions, iii.number of errors and iv.ratio of correct and incorrect answers Analysis of the user testing found that no particular interface configuration exhibited increased efficiency, effectiveness, learnability or satisfaction and that users were able to adapt to a changed interface with ease once they had become accustomed to the word processor environment. Therefore, the final finding of the study was that provision of an interface in a bilingual userâs first language neither significantly contributed nor detracted from the applicationâs usability. Similarly, neither of the pictorial icon sets nor the text buttons exhibited a significantly heightened level of usability. Therefore, none of the interface configurations could be recommended as the most usable. However, a number of recommendations concerning the usability of a word processor were proposed based on both the analysis of the tasks and observation of user interaction. Finally, based on user performance for each individual task, an icon was identified which appeared to be the best and most applicable for that function. The final recommended interface, the usability of which must still be empirically established, consisted of a combination of standard icons, alternative icons and text buttons.

Not available

The AIDS epidemic, caused by Human Immunodeficiency Virus (HIV), is a global crisis which threatens development gains, economies, and societies. The eradication of HIV/AIDS represents one of humanityâs greatest challenges, which requires co-operation, and comprehensive collaboration between science, governments, social institutions, the media, the professions, and the general public. In this endeavour strategic information plays a major role. The Free State Department of Health introduced Antiretroviral Therapy (ART) to meet the needs of the rising number of people living with HIV and also dying due to AIDS related diseases. A patient information system was deployed by the Province to supplement the ART programme rollout process by gathering data and providing all the basic patient antiretroviral information. The patient information system was a traditional online clinical system and concentrated on the bread-and-butter issues of accumulating data on a patient. Very little functionality was provided to deal with the complexities of managing the clinical outcomes of the ART programme. To add to the problem, other online operational systems had to be interrogated to gain an understanding of the impact the rollout of ARV had. These operational systems ranged from standalone Human Resource systems to information systems accumulating data on tuberculosis which is closely related to HIV/AIDS. No mechanism or platform existed to provide management with integrated strategic information to manage the business process intelligently. This study focussed on the challenges and solutions to overcome this shortfall. The main objective of this study was to construct a framework for providing strategic information for the management of the ART programme in the Free State Department of Health. This objective was reached with a two phase action research methodology. The author of this study fulfilled simultaneously the roles of the researcher and the practitioner. During phase one a central data warehouse was designed and developed that incorporated an existing standalone ARV data warehouse and several other independent operational sources, all related to ART. This warehouse was evaluated by the users who overwhelmingly rated the warehouse as successful. From one system it was possible for managers to obtain strategic information on ARV encounters, ARV human resources, revenue collection, in-patient hospitalization, notifiable diseases and tuberculosis. This was achieved with a very limited budget and using internal resource. Although the data warehouse was a valuable asset by itself, several improvements were identified during the evaluation of the warehouse. The most important of these were to add additional data marts, to conform all the dimensions in order to obtain strategic information across the data marts and lastly to be able to track an ARV patient over time over all facilities of care. This led to the second phase of the research. In this phase several data marts (National Health Laboratory Services (NHLS), Meditech hospitalization and linkage to Home Affairs) were firstly added. Secondly the patient dimensions of all the data marts were conformed through the process of probabilistic record linkage. Lastly a longitudinal patient record was developed that displayed all the encounters of an ARV patient over time. Even though the LPR could not be scientifically evaluated, the institutions that did test it rated it very highly and reckoned it was a first for South Africa. This study, therefore, concluded with the fact that it was possible to construct and implement a successful framework for providing comprehensive strategic information for the management of the ART programme in the Free State Department of Health.

In recent times, people have become more and more reliant on computers on a daily basis. As a result, the need has arisen to optimise the task-related experience in terms of time efficiency, which demands effective training in software skills. To be more specific, word processing skills are currently considered essential in any field of work and are in high demand. This study focuses on determining the optimal paradigm (methods) to assess usersâ word processing skills. One of the main reasons for this research was the fact that students at the University of the Free State (UFS) reported to the computer literacy course lecturer that they were dissatisfied with the virtual, simulated MS Word software environment used to assess (e-assess) their word processing skills electronically. This existing test system (ETS) at the UFS requires students to perform certain tasks and automatically checks whether the required end-result is obtained. However, this system is based on a simulated interface with limited functionality. As a result, the relevant information on software e-assessment systems was researched and a new software skills e-assessment application developed accordingly. The aim was to develop a tool that would be able to assess studentsâ word processing skills in the most reliable way possible. Another aim was to find methods of stimulating the learning process during the e-assessment of word processing skills. Therefore the newly developed e-assessment system, WordAssessor, is designed to be based on the real MS Word environment. It requires students to perform certain tasks and automatically checks whether the said tasks have been correctly performed. WordAssessor allows students to explore the MS Word interface fully while being assessed. It even allows students to use trial-and-error to solve certain problems (tasks). To stimulate learning potentially further during e-assessment, WordAssessor presents students with a video solution for the questions they had incorrect, directly after the test. In order to assess the validity of the e-assessment methods employed by the WordAssessor system, comparative user testing was conducted. Studentsâ word processing skills were assessed as part of their advanced computer literacy course, using the ETS, the WordAssessor system and a personalised test scenario (where no e-assessment tool was used). In addition, participating students were provided with a questionnaire to determine their reaction and preference with regard to the various elements of the assessment methods. By analysing the results in detail, it was found that the results of the personalised test scenario (PT) yielded the most reliable indication of studentsâ true word processing skills, and could be used as a benchmark. Hereafter, the results of the WordAssessor test scenario were analysed to determine the correlation (relationship) with the results of the personalised test. The same type of correlation was performed between the results of the ETS and PT. It was established that the WordAssessor results correlated significantly more closely with the PT results than the ETS. In the end (and after additional analysis) it was found that the methods employed by WordAssessor yielded the most reliable indication of studentsâ true word processing skills knowledge when compared to the ETS. In addition, from the results of the post-assessment questionnaire, it was determined that students felt they learned more as a result of the video tutorials. Furthermore, they stated that they preferred video feedback over text- or paper-based feedback. They also stated that they preferred being assessed in a real software environment, as opposed to a simulation. It was recommended that a more flexible and realistic e-assessment approach (as demonstrated by the WordAssessor system) could be beneficial to students on several levels. Also, it was recommended that students be able to answer test questions in a way that suits them rather than being instructed as to the method of use. Finally, the use of highly detailed video tutorials directly following e-assessment (only for questions students had incorrect), was recommended.