Artificial Intelligence for Medical Image Analysis


Segmentation can be useful in a range of clinical applications, such as biomarker estimation and treatment planning. The group is active in this area, proposing novel nut clinically applicable methods for using AI to automated challenging segmentation tasks.

Topology-informed segmentation

There are many cases in which the topology of a structure to be segmented is known a priori. For example, when segmenting the left ventricle of the heart from short-axis cine MR images, it is known that the correct segmentation will be a topological closed loop in two dimensions. However, traditional deep learning based segmentation models make no use of this knowledge. In this work, we explicitly incorporate topological priors into a deep learning segmentation model for the first time, based on the concept of persistent homology. These priors require no additional ground truth segmentations and so the model is suitable for use in a semi-supervised setting. We found that by including the topological priors the number of topological errors was reduced whilst maintaining high overlap with the ground truth. The initial work demonstrated this principle for single-class segmentation but we subsequently extended it to the multi-class case.


Examples of a topology-informed loss function fixing topological errors in short axis cardiac MR segmentation.

Automated assessment of spleen size from ultrasound

Examples of automated spleen segmentation and length measurement.

Patients suffering from Sickle Cell Disease can suffer from a range of complications including abnormal enlargement of the spleen. Ultrasound forms part of the clinical pipeline for spleen assessment, but measuring its size in ultrasound images is a subjective process. Furthermore, in many parts of the developing world there is a lack of skills for performing reliable spleen assessments. We investigated the use of deep learning to automate this process, reaching almost human-level performance.


Segmentation of the Vocal Tract From Speech MRI

MR is increasingly playing a role in speech analyis - dynamic MR images of patients acquired during speech can be used for planning and monitoring of treatment for speech problems such as velopharyngeal insufficiency. In this project our ultimate aim is to construct dynamic 3D speech models of patients for informing treatment decisions. Initial work has investigated the use of deep learning for full segmentation of the articulators in the vocal tract from 2D MR, resulting in the first fully-automatic model demonstrated for this task. The dataset and code for training this model has been publicly released.



Automated segmentation of the vocal tract articulators from dynamic MR images: left - ground truth, middle - deep learning based segmentation, right - after post-processing.