Artificial intelligence in the form of deep learning, for instance using convolutional neural networks, has made a huge impact on medical image analysis. It dominates conference and journal publications and has demonstrated state-of-the-art performance in many benchmarks and applications, outperforming human observers in some situations. But, despite this, adoption of these approaches in routine clinical practice has been very slow. Indeed, while they can present extraordinary performance in some tasks, they can also fail spectacularly making mistakes that a human observer would consider non-sensical. One reason for this is that deep learning models for imaging analysis often tend to consider only one type of information when making their prediction – the image. However, understanding and interpretation on what is on an image by a clinician often involves external knowledge, such as standard anatomy, how structures are deformed due to the presence of a lesion (the shape of a bone is less likely to be deformed by a tumor than a soft tissue), knowledge of how a specific disease evolves or changes the appearance of the affected tissue but also of its surroundings etc. The project will thus investigate approaches to condition neural networks using this expert knowledge to both improve prediction in multiple tasks but also limits catastrophic failures making these predictions more reliable and thus more suitable to be implemented in clinical practice. This allows for the development of clinical applications that would help the radiologist make informed and precise decisions in less time and could also drive the development of screening tools to flag at-risk patients to clinicians.
CDT Essential Criteria
A Masters level degree (MEng, MPhys, MSc) at 2.1 or equivalent.
Desire to work collegiately, be involved in outreach, undertake taught and professional skills study.
Project Essential Criteria
Experience in machine learning/deep learning.
Experience in image processing/computer vision.
Project Desirable Criteria
An interest in healthcare/medical imaging.
The CDT in Applied Photonics provides a supportive, collaborative environment which values inclusivity and is committed to creating and sustaining a positive and supportive environment for all our applicants, students, and staff. For further information, please see our ED&I statement https://bit.ly/3gXrcwg. Forming a supportive cohort is an important part of the programme and our students take part in various professional skills workshops, including Responsible Research and Innovation workshops and attend Outreach Training.