University of Cambridge Home Department of Engineering

F-AHG-1: Modelling the human semicircular canals

Benign paroxysmal positional vertigo (BPPV) is a bothersome, often long-standing condition that can be relieved through the particle repositioning manoeuvre (PRM). This involves rotating the patient in the plane of the posterior semicircular canal, to displace free floating particles that are trapped in the canal and are responsible for the vertigo.

This project will develop tools to assist the clinician locate the plane of the posterior canal as they perform PRM. We will start with a large collection of CT scans of the human inner ear, from which we will extract the semicircular canals using an exiting model-based segmentation method. We will also locate anatomical landmarks on the outside of the head (e.g. corners of the eye sockets, external ear canal, nasion) that could be used as points of reference during PRM.

Next, we will develop statistical models to capture the inter-subject variation in the anatomy. These models will help us answer questions like "How much does the orientation of the canals' planes vary with respect to externally identifiable landmarks?" And "What are the best set of external landmarks to use, so as to locate the canals' planes?"

Finally, we will consider how best to convey the resulting information to the clinician during PRM. Possibilities would be some sort of 3D printed frame to attach to the external landmarks, or computer vision tracking of the head driving a real time display of the canals.

The project is offered in collaboration with Professor Manohar Bance and Dr Iwan Roberts at Addenbrooke's Hospital. The project would suit a student who has taken Module 3G4 and Project GG2, though neither is a strict prerequisite. It will involve programming in Python and/or C++, and the opportunity to learn more about medical image analysis and graphics. See this paper for previous work we have done on inner ear imaging.

Model-based segmentation of the human cochlea and semicircular canals in clinical CT images. A large collection of CT scans will be segmented in this manner, including additionally anatomical landmarks on the outside of the head that could be used as points of reference during PRM.
This animation shows the first mode of variation in a statistical shape model of the otic capsule (the cochlea and the semicircular canals). This project will develop new models focusing just on the semicircular canals and also the external anatomical landmarks.
© Cambridge University Engineering Department
Information provided by Andrew Gee
Last updated: April 2023