Alex Harston

PhD Student

Alex is a third year PhD student in the lab, funded by the Department of Bioengineering. His research focuses on using machine learning techniques to build a data-driven understanding of real-world human attention. He has a BSc in Neuroscience from the University of Edinburgh, and worked as a research assistant & developer prior to starting his PhD. He works as a web developer in his free time, and built this website.

Research Brief

Alex works on developing data-driven methods to understand human attention in the real-world. His research interests lie in understanding the links between perception and action in real-world tasks, with a focus on developing computational approaches to predicting real-world visual saliency in natural tasks.

Publication List

Linear Embodied Saliency: a Model of Full-Body Kinematics-based Visual Attention


Abbott, William W and Harston, J Alex and Faisal, A Aldo


Human visual attention prediction boosts learning & performance of autonomous driving agents


Makrigiorgos, Alexander and Shafti, Ali and Harston, Alex and Gerard, Julien and Faisal, A Aldo

arXiv preprint arXiv:1909.05003

How body movements in a task predict visual attention dynamically


Harston, John A and Abbott, William W and Faisal, Aldo

The Association for Research in Vision and Ophthalmology

The Embodied Semantic Fovea: Real-time understanding of what and how we look at things in-the-wild


Faisal, Aldo A and Harston, John A and Auepanwiriyakul, Chaiyawan and Li, Mickey and Orlov, Pavel

The Association for Research in Vision and Ophthalmology

Generalised Structural CNNs (SCNNs) for time series data with arbitrary graph topology


Teh, Thomas and Auepanwiriyakul, Chaiyawan and Harston, John Alexander and Faisal, A Aldo

arXiv preprint arXiv:1803.05419

Semantic fovea: real-time annotation of ego-centric videos with gaze context


Auepanwiriyakul, Chaiyawan and Harston, Alex and Orlov, Pavel and Shafti, Ali and Faisal, A Aldo

Proceedings of the 2018 ACM Symposium on Eye Tracking Research \& Applications