Ultrasound Nerve-block Guidance using Machine Learning

One of the most critical components of ultrasound-guided nerve blocks is the accurate detection of a nerve within an ultrasound image. Trying to classify the various types of tissue in a noisy, greyscale ultrasound feed is a very daunting task. Therefore, a software solution would help doctors to perform the procedure in a precise and effective manner. To date, there have not been many improvements in terms of nerve detection, often resulting in the application of more anesthetic than required. This is problematic because doctors aim to use the minimum amount of anesthetic which poses less of a risk for the patient’s health. This study will present a software-based solution for the detection of nerves in ultrasound images. The software will take advantage of machine learning and image processing techniques to detect the location of the various nerves in the body within an ultrasound feed.

Faculty Supervisor:

Thomas Hemmerling

Student:

Partner:

Divocco Medical

Discipline:

Life Sciences

Sector:

Retail trade

University:

Research Institute of the McGill University Health Centre

Program: