Some of the research being conducted by members of the IVES research group
LOW-COST COMPUTER-BASED REHABILITATION SYSTEM FOR STROKE SURVIVORS
Lead Investigator: Dr Mohd Fairuz Shiratuddin
This research focuses on designing, developing and evaluating a low-cost computer-based stroke rehabilitation system for stroke survivors. Some of the sub-research areas involved in this research include data analytics & data science: data mining & big data, machine learning, human-computer interaction and natural user interfaces, virtual environments and game technologies
The research is still on-going, and we hope to do a pilot trial soon.
As part of the research, a prototype system called Neuromender has been developed to support our study and data collection.
The system is developed by our Computer Science, Games Technology, and Games Software Design & Production students.
Neuromender comprises of hardware and software. The software is developed using the Unity 3D Game Engine for the 3D virtual environments, game mechanics, artifical intelligent, game logic, and game-based activities. The software also has an online backend data collection portion which is written in PhP, and using MySQL database. Neuromender utilises different types of Natural User Interfaces (NUIs) hardware devices which include Microsoft Kinect 1.0 & 2.0 for Windows, Razer Hydra and Leap Motion. Beside using a typical computer monitor, Neuromender also supports Oculus Rift DK1 & DK2 Head Mounted Display (HMD) to display content in the 3D virtual environment for a more immersive experience.
Since we are targeting Neuromender for stroke survivors who have either very limited control or gained good control of their upper-limb motor control, the game-based acitivities are designed to accomodate this.
FLEXIBRAINS
Lead Investigator: Mr Shri Rai
FlexiBrains is a web-based cognitive retraining suite developed for stroke survivors. Its goal is to present challenging visual, visuospatial and aural tasks to promote the rehabilitation of stroke clients. FlexiBrains monitors the accuracy and latency of clients when performing individual tasks and presents the data to clinicians for assessment. It has an intuitive user interface and is accessible through a computer with the latest web browser as well as any modern tablet with Internet connection.
A PILOT STUDY IN MEASURING USERS’ TASKS PERFORMANCE USING NATURAL USER INTERFACES (NUIS) IN AN INTERACTIVE REAL-TIME 3D VIRTUAL ENVIRONMENT
Lead Investigator: Dr Mohd Fairuz Shiratuddin
This project aims to design and develop suitable tasks and then measure the users’ task performances when exposed to different types of NUIs in an interactive 3DVE. In this project, the types of NUIs that will be tested include; contact/touch and non-contact/non-touch gestures (either finger/s, hand/s and/or whole body). The outcomes of this pilot project will assist in in the establishment of a framework of NUIs in a 3DVE that can be utilised for the rehabilitation of stroke patients’ motor control.
MURDOCH DRIVING SIMULATOR
Lead Investigator: Mr Shri Rai
The driving simulator system uses a client-server model where multiple drivers (clients) are connected to a central server running a virtual world. The system permits the immersion of drivers in various driving enviroments. Road environments can be interactively constructed within the simulation.A special client on a control console can select and dynamically adjust vehicle or environment parameters of client vehicles and/or world environment whilst the simulation is being run. Drivers sit in custom driving rigs and all interactions of the driver is monitored.
Physiological monitoring is also carried out to determine stress levels of the driver to various driving scenarios during the simulation run.
We are collecting detailed driving performance data in order to characterise differences in behaviour and physiological responses in response to challenging driving environments as a function of their prior driving experience. On completion of this phase of the project, we would like to to be able to put anyone into the simulation and determine the amount of driving experience they may have and consequently determine situations in which an individual might be involved in a real traffic incident. There are a number of application areas and one of these areas would be the use of this information to provide a better targeted education to drivers of the hazards on the road. It would also enable us to have a better understanding of how dangerous situations arise during driving and the length and type of driving experience needed to reduce and hopefully prevent dangerous accidents. Drivers cannot be made to experience dangerous situations on a real road because of the obvious consequences. Doing it on a test track is expansive and not available to everyone. It would also involve damage to vehicles. Simulators are a cost-effective way to provide training to drivers so that they learn how to handle difficult situations.
The driving simulator system is one of a number of simulation systems built in-house over a number of iterations and it provides practical software engineering training to our undergraduate games technology students. The system was written from the "ground-up" using C++. As the system was built in-house, we have control of system specifications and the development of the system. Our graduating games technology students also end up having a more in-depth experience of software maintenance. The system is then used by us and our research students as a research platform.
Initial funding for the simulator came from the state of Western Australia road safety council.
RADIOLOGY ASSISTANCE TRAINING SIMULATOR (RATS)
Lead Investigator: Dr Hong Xie
RATS is a simulator to train students on how to use radiology equipment such as an X-Ray machine. RATS encompases a real-time 3D virtual environment to simulate the actual environment itself. RATS provides a much safer option as it could avoid accidental exposure to radiation, it is low-cost and can be run on any average modern PCs. This project is still ongoing and will soon include the use of a Head Mounted Display to provide a much more immersive experience to the trainee.
AGED CARE USING VIRTUAL ENVIRONMENT
Lead Investigator: Associate Professor Dr Kevin Wong
The impact of Dementia on elderly is increasing in a fast pace in the recent years. Normally, exercise and diet are both factors that can help to prevent this problem. Cognitive activity is also another important factor to consider. While exercise and diet can be addressed by individual, cognitive activity for those individuals that involves some sort of social engagement can be difficult to find. Bingo, especially when played with peers, is commonly cited as an activity that can assist with training memory and information processing. For some seniors who are not mobile, it will be a problem for them to go somewhere and participate in a Bingo game. This project aim to look for an alternative using virtual environment and natural user interface. This project also examine ways to track and perform data analytics on the progress of the senior in such cognitive activity.
EMOTION RECOGNITION USING EEG
Lead Investigator: Associate Professor Dr Kevin Wong
Human Emotion is an important area of study in the Human Computer Interfaces (HCI) discipline. In this project, methods used to enhance the emotion recognition are investigated and developed. Efficient ways of handling emotion recognition or emotion classification is becoming more and more important with the advancement of Brain Computer Interface (BCI) in many applications. This project is focusing more on real time emotion recognition using EEG.
DATA MINING WITH IMBALANCED DATA
Lead Investigator: Associate Professor Dr Kevin Wong
The class imbalance problem is one of the important problem exist in many data. An imbalanced data set could be one of the obstacles for several machine learning algorithms. Many existing algorithms cannot be used directly to handle the imbalanced data problem especially in multi-class classification. This project investigates ways and strategies to handle the imbalanced problem in many current data analytic problems.
FUZZY RULE INTERPOLATION
Lead Investigator: Associate Professor Dr Kevin Wong
Fuzzy logic can be found in many applications ranging from domestic products like washing machine to many commercial control systems. However, the nature of data for current applications may pose many challenges for current fuzzy system. Fuzzy rule interpolation (FRI) techniques were introduced to generate inference for sparse fuzzy rule base to perform inference with a reduced set of fuzzy rules. Basically, FRI techniques perform interpolative approximate reasoning by taking into consideration the existing fuzzy rules for cases where there is no fuzzy rules to fire. See http://fri.gamf.hu/
FUZZY SIGNATURE AND COGNITIVE MODELLING
Lead Investigator: Associate Professor Dr Kevin Wong
As data is getting more complex and complicated, it is increasingly difficult to construct an effective complex decision model. Fuzzy signatures are introduced to handle complex structured data and problems with interdependent features. A fuzzy signature can also be used in cases where data is missing. Fuzzy signatures can address some issues of granulation and organisation well. In order to better model the human cognitive system, we have divided our cognitive modelling into two main categories. The first category consists of meta-levels of visual representation to model decision and cognitive behaviour. In this category the model consists of nodes and pointers to show the concepts and relations. Each node exhibits the behaviour of a human cognitive system. Each node consists of three states, the sensory input state, current state, and action state. In the second category, nodes basically consist of the fuzzy signatures. These signatures contain the knowledge necessary for the node to take any action.
EMERGENCY EVACUATION SIMULATOR
Lead Investigator: Mr Shri Rai
FORMATION OF THE INTERNATIONAL WOMEN’S EDUCATIONAL LEADERSHIP FORUM, AN ORGANISATION WHICH AIMS TO PROVIDE TARGETED MENTORING FOR YOUNG WOMEN AND THOUGHT LEADERSHIP FOR GLOBAL EDUCATION
Lead Investigator: Professor Sara de Freitas
SCIENCE OF LEARNING PROJECT TO INVESTIGATE EFFICACY OF LEARNING IN DIFFERENT MODES THROUGH NEUROPSYCHOLOGY TECHNIQUES USING EEGS
Lead Investigator: Professor Sara de Freitas
COMPUTER ANALYSIS OF EEG FOR DYSLEXIA DETECTION
PhD Student Researcher: Ms Harshani Perera
Supervisors: Dr Mohd Fairuz Shiratuddin, Associate Professor Dr Kevin Wong
Dyslexia, a disability that causes lack of proficiency in reading, writing or spelling despite normal (or above) intelligence and sensory abilities is found among approximately 15 to 20 of percent of the global population. Although dyslexia, commonly know as ‘word-blindness’ has been known for the last ten decades and is spread among a significant number of the population, unfortunately it still often goes undetected.
This research focuses on computer analysis of Electroencephalogram (EEG) to detect dyslexia. An improved framework for analysis and classification between dyslexics and non-dyslexics will be introduced. Oscillations of brain electric potential are analyzed in dyslexics when performing different tasks and compared to non-dyslexics (normal) to identify unique patterns in the EEG band frequencies. Dyslexia detection using EEG would be more reliable since the brain wave outcome cannot be falsified unlike the conventional behavioral based dyslexia detection techniques.
AN INVESTIGATION OF THE EFFECTIVENESS OF LEARNING USING MOBILE AUGMENTED REALITY (mAR)
PhD Student Researcher: Ms Siti Salmi Jamali
Supervisors: Dr Mohd Fairuz Shiratuddin, Associate Professor Dr Kevin Wong, Dr Charlotte Oskam
This research supports the student centred learning concept through the use of mobile-Augmented Reality (mAR) as assistive learning tool. Kapp and McAleer (2011) state that mAR will be the most intriguing medium in learning initiative and could enhance learning especially for science students. The main aim of this research is to investigate the effectiveness of learning using mAR.
PERSUASIVE VISUAL DESIGN FRAMEWORK FOR WEBSITE DESIGN
PhD Student Researcher: Mrs Nurulhuda Ibrahim
Supervisors: Dr Mohd Fairuz Shiratuddin, Associate Professor Dr Kevin Wong
In an online environment, users are constantly engaged to make choices and decisions. However, due to the nature of the Internet of being massively loaded with so much information, making the correct decision is difficult especially when users have hard times looking for the relevant and required information (Chen, Shang, & Kao, 2009). One of the main challenges for businesses and web designers is to convince web visitors or users to stay at their site, and hope to succeed in influencing them to take actions that can benefit both parties (Zhang & von Dran, 2000). However, convincing users to stay at a particular website is not an easy task and with many other similar websites available, can easily steer them away.The website’s design must be able to attract users’ attention so that they are motivated to make the decision to stay. Designing for persuasion is a possible way to attract users’ attention. Persuasion affects motivation (Fogg, 2003) and persuasion itself is about making a choice (Larson, 2010). However, not much guidance can be found for creating attractive website visual designs (Sutcliffe, 2001b). Sutcliffe adds that experts merely give advice in the form of abstract and general laws, which in turn leave web designers to interpret the design cues on their own.
The aim of the research is to develop a framework of persuasive visual design for website design. This research seeks to learn what visual design factors that will influence users in making the decision whether to stay or leave a site and if they decided to stay, will the design elements encourage them to make further actions. This research will also investigate in what way visual design can influence the users and what are the impacts of persuasive visual designs on users’ motivation. The triggers and barriers in optimising users’ motivation will be identified. Even though the expected framework can be generalised to various contents; however, this research will be based on the subject of tourism.
USING ROUGH SET THEORY TO IMPROVE CONTENT BASED IMAGE RETRIEVAL SYSTEM
PhD Student Researcher: Mrs Maryam Shahabi
Supervisors: Dr Mohd Fairuz Shiratuddin, Associate Professor Dr Kevin Wong
Each image in a Content Based Image Retrieval (CBIR) system is represented by its features such as colour features, texture features and shape features. These three groups of features are stored in the feature vector. Therefore each image managed by the CBIR system is associated with one or more feature vectors (Lee, Kim, and Kim 2012). As a result, the storage space required for feature vectors is proportional to the amount of images in the database. In addition, when comparing these feature vectors, the CBIR system understand which images in the database are similar to another (Li, Fan, et al. 2012).
Nonetheless, researchers are still facing problems when working with huge image database since so much time is spent to compare huge feature vectors that require large amount of memory to run the CBIR system. Due to this problem, feature reduction and selection techniques are employed to alleviate the storage and time requirements of large feature vectors. There are many feature reduction techniques, including linear projection techniques such as the Principal Component Analysis (PCA) and the Linear Discriminate Analysis (LDA), and the metric embedding techniques (both linear and non-linear) (D.DeMers and G.Cottrell 1993). However, these methods have limitations in the CBIR system and could not improve CBIR performance (retrieval accuracy) and reduce semantic gap efficiently. Therefore, we need a feature selection method that can deal with image features efficiently and has the ability to deal with uncertainties.
In this research, the author proposes a new pre-processing stage to overcome these problems and improve the CBIR system. This pre-processing stage selects important features from a massive image feature vectors, omitting features, which are vague, uncertain and not important. These redundant and vague features will influence the analysis by generating undesired retrieval. Consequently, from these significant features, semantic rules are then extracted that can classify images more accurately and show more relevance images to the user, hence improving the retrieval accuracy.
FRAMEWORK FOR VISUALISING MUSIC MOOD USING TEXTURE
PhD Student Researcher: Mrs Adzira Husain
Supervisors: Dr Mohd Fairuz Shiratuddin, Associate Professor Dr Kevin Wong
The traditional way of browsing music collection is by going through a textual list of song, and the searching methods may not be sufficient to maintain an overview of the collection (Torrens, Hertzog, & Arcos, 2004). Instead of using textual lists, another way to browse music collections is by using some type of visual form. In the area of music information retrieval, many researchers have visualised music collection in various ways to make music browsing more interesting and efficient; this includes tree maps (Torrens et al., 2004), geographical maps (Kornhauser, Wilensky, & Rand, 2009; Leitich & Topf, 2007), rainbows (Pampalk & Goto, 2006), 3D spiral (Lamere & Eck, 2007) and many more.
Recent studies include finding the suitable visual variables for musical metadata such as tempo, genre and release year (Holm & Siirtola, 2012), and browsing music collections based on mood (Lehtiniemi & Holm, 2012; Schatter & Kramer, 2011). There are also a number of research on music visualisation using visual variables such as position, size, shape, value and colour but there is no research focusing explicitly on textures (Holm, 2012). Research shows that texture is an associative visual variable (Carpendale, 2003), but there is still no prove that texture can be associated with mood.
Thus, the main aim of this research is to develop a framework for visualising music mood using texture. This will require an investigation on how texture image can represent music mood that can easily be identified and produce a perceptual impact to the user.
A CROSS-DISCIPLINARY AND EVIDENCE-BASED STUDY OF HUMAN LEARNING
Lead Investigator: Dr Victor Alvarez
MORE RESEARCH DESCRIPTIONS BY OUR RESEARCH STUDENTS ARE COMING SOON
Mr Ali Elkaseh (PhD)
Mr Ibraheem Al-Jadir (PhD)
Mr Moamer Ali Ahmed Shakroum (PhD)
Mrs Shalini Christobel (PhD)
Mrs Asma Smeda (PhD)
Mr Abdughani Abeid (PhD)
Ms Jamie Shih Chi Tang (PhD)
Ms Mehrnaz Akbari Roumani (PhD)
Mr Varin Khera (DIT)
Mr Pema Choejey (PhD)
Mr Wisut Chamsa-ard (PhD)
Mr Anuchin Chatchinarat (PhD)
Miss Ratchakoon Pruengkarn (PhD)
Mr Moamer Shakroum (PhD)
Mr Sanal Panicker (PhD)
Mr Robert Herne (Honours)