Lecturer

• Dr Ibrahim Khalil

Papers

Research questions

1. In mobile healthcare enormous amount of ECG data needs to transferred from patient mobile devices to a centralized server in a medical centre. Therefore, it makes sense to compress this data. Why conventional compression techniques are not very efficient here? How should this data be compressed?
2. The Health Insurance Portability and Accountability Act (HIPAA) requires that patient medical data (e.g ECG) be transferred securely from medical sensors (using bluetooth) to mobile PDAs (i.e. data collecting devices). How data can be securely collected by the PDAs?
3. This is similar to Q2. The Health Insurance Portability and Accountability Act (HIPAA) requires that patient medical data (e.g ECG) be transferred securely from  mobile PDAs to a server in a medical centre. How can ECG data be securely and efficiently transmitted  over public  internet?
4. "The ECG may roughly be divided into the phases of depolarization and repolarization of the muscle fibres making up the heart. The depolarization phases correspond to the P-wave (atrial depolarization) and QRS-wave (ventricles depolarization). The repolarization phases correspond to the T-wave and U-wave (ventricular repolarization)." Using these ECG features, a person can be uniquely identified. How can we anonymize  the data to prevent malicious users from gaining access to a patient's cardiovascular condition?
5. A patient database system contains records of large number of patients. Only selected medical personnel in a medical centre should be allowed access to selected medical records? How can you ensure that?
6. In mobile healthcare, a mobile PDA is required to measure the distance of PR interval from the given ECG data. A normal PR interval is 0.12 - 0.20 seconds. A short PR indicates "white syndrome" while a long PR indicates "first/second degree AV block". Provide a simple algorithm to calculate PR interval from given ECG data. Compare your proposed algorithm with existing solutions. (note: we will provide raw ECG data).
7. This is similar to Q6. Provide a simple algorithm to calculate QRS interval from given ECG data. Abnormal QRS indicates right/left bundle branch block.  Compare  your proposed algorithm  with existing solutions.
8. Again, this is similar to Q6. Provide a simple algorithm to calculate QT interval from given ECG data. Abnormal QT indicates coronary disease, myocardial disease etc.  Compare  your proposed algorithm  with existing solutions.
9. Localization, detection delay, and resolution that  satellite-based wildfire detection techniques  provide is not  adequate for all cases. Why? Compare the pitfalls of different techniques.
10. In sensor-based bushfire monitoring high temperature can cause the system to send alarm. However, "high-temperature objects such as engines or chimneys that aren’t burning usually cause  false alarms." What needs to be done to prevent this? Compare existing solutions and propose a new one.