Andrew Broad
- Email: scajb@leeds.ac.uk
- Thesis title: Can Attention-Inspired Artificial Intelligence Provide a Diagnostic Understanding of Colorectal Cancer Imaging Data?
- Supervisor: Dr Marc de Kamps, Dr Alex Wright
Profile
I took my first degree at Leeds University, in Electrical and Electronic Engineering. I then worked as a broadcast engineer at BBC Leeds for a while, before moving into software development. This began at Bradford University, where I wrote music synthesis code for an electronic organ project. More recently I have found it most rewarding to work on NHS projects, such as the NHS Choices website, data acquisition systems at the Information Centre (now part of NHS Digital), and ultimately the PPM+ electronic health record at Leeds Teaching Hospitals NHS Trust.
I am now studying for an integrated MSc/PhD in Artificial Intelligence, in a Centre for Doctoral Training (CDT) programme involving the University and Leeds Teaching Hospitals.
Research interests
I am part of a research team involving the School of Computing and the National Pathology Imaging Cooperative. We are investigating how AI can be used in histopathology, to help diagnose cancer by identifying cell types in digital whole-slide images (WSIs) of human tissue.
The central problem here is that of scale; the scanned image can be tens of gigapixels in size, and would cover the area of a tennis court if printed out at typical resolutions. Popular AI algorithms can only analyse relatively tiny image patches. For acceptable speed, and to reject areas of noise, we need to avoid processing the entire WSI patch by patch. For this, our AI processing pipeline has to decide where to look.
I am investigating approaches based on human attention, where we naturally pick out areas of interest to build up an understanding of a complex scene. Early prototypes have shown promising results in outlining the tumour region, and in calculating cell ratios that can predict a cancer patient’s response to different treatments. Future work will build on this by incorporating the latest findings from computational neuroscience, to make our system more accurate and efficient.
Qualifications
- B.Eng (Hons), Class I, Electrical and Electronic Engineering, University of Leeds
- BBC Engineering Training Standing Instructions (ETSI)
- C.Eng Chartered Engineer