Project # 1
Emotion conveys our cognitive state, guides our behavior, and influences our motivation. Users of psychoactive drugs (e.g., amphetamine, cannabis, and opioid) usually show some emotional impairment as a result of brain damages caused by these illicit drugs. In Malaysia, ketum has been widely used as psychoactive natural substance for its medical and euphoric effects. Although the use of ketum is reported to cause unpleasant psychological withdrawal symptoms in regular ketum users, to date, information on emotional impairment among ketum users has not been explored and scientifically investigated.
In general, there are different methods that are used for detecting the emotional impairment such as EEG, fMRI, and EMG. However, all of these equipment are costly, require constant expert monitoring, and more importantly, obstrusive. Prior studies demonstrated that gaze movements could reveal how people process emotional cues such as facial expressions. Hence, this project aims at exploring an alternative automatic seamless and unobtrusive measuring tool based on analyzing the gaze movements of ketum users. Advances in deep machine learning is going to be utilized to detect, track and analyze ketum users and non-users to establish an automatic and reliable model to detect the emotional impairment in ketum-users.
This project will be performed in close collaboration with drug center experts and postgraduate researchers. The applicant would be required to review the recent studies about emotion processing in psychology and emotion detection in computer science, collect data from ketum users, implement techniques for gaze detection and publishing.
Project Duration: 1 year
Starting: Jan 2016 (but can already start with literature review)
Project # 2
Cyanobacteria or the blue green algae are known for their capability to produce toxic secondary metabolites termed cyanotoxins. These toxins are responsible for human health problems and animal poisonings worldwide, ranging from skin irritation to more harmful effects such as organ failure. Cyanobacteria can be confused with true algae or water weeds and this confusion may have obscured early detection of its toxic bloom. Early detection of toxic bloom requires rapid identification of cyanobacteria in water resources. By far, phenotypic evaluation is the simplest approach proven to be useful for instant detection of the presence of toxic cyanobacteria. However, lack of knowledge in cyanobacteria taxonomy, as well as locally available expertise, and lack of samples to be identified often led to the misidentification of this type of algae. This warrants for a method to automatically identify of cyanobacteria using an array of available but limited features.
This interesting project will be carried out under the collaboration with the School of Biological Sciences. The candidate will have a field supervision from an algae expert.
The student is to explore image processing and deep learning and develop a classification system of Cyanobacteria – that would tell Cyanobacteria from other true algae. Only offline modelling is expected, and not online testing (real-time testing). Data will be provided.
JAN 17- SEPT 17 (the prospective candidate can already start before that, if he/she is ready)