Cardiac diseases are one of the most prominent diseases world-wide Invasive cardiac procedures are required to diagnose non-working heart tissue (e.g., cardiac blocks) before surgery. These invasive procedures themselves might cause strokes, heart attacks, etc. Correctly identifying the exact heart tissue that needs to be operated/removed to rectify the cardiac ailment without invasive procedures would be a significant medical breakthrough.
Our research group has had initial success in correctly identifying heart tissue ailments from body surface potentials without the need for invasive diagnostics using neural networks. In this project students will develop further on our state of the art neural network designs with the aim to correctly diagnose heart surface potentials from body surface potentials.
A neural network capable of predicting the heart surface potential from body surface potentials thereby enabling non-invasive clinical diagnostics.
Interest in machine learning.
Lab allocations have not been finalised