The aim of this study is to determine if machine learning can be used to automatically highlight key anatomy on the ultrasound image to help anaesthetists perform ultrasound-guided regional anaesthesia.
The study will involve adult volunteers who are willing to be scanned by a trained sonographer to collect ultrasound video data for the following categories: * Adductor canal * Popliteal fossa * Fascia Iliaca * Rectus sheath * Axillary * ESP * PECS * Interscalene Each volunteer will be scanned to collect data for every category in the list. Where applicable, both sides of the body will be scanned. The videos will be segmented by hand to identify the relevant anatomical regions for each category. The primary objective for this study is to provide the range of data required to develop robust models in conjunction with additional data from patients undergoing a Peripheral Nerve Block procedure that are able to produce the desired segmentation on the unseen validation images. The models will be scored using the standard "Mean intersection over Union" pixel-level metric for semantic segmentation.
Study Type
OBSERVATIONAL
Enrollment
30
Medicentre
Cardiff, United Kingdom
Model development
Models with a Mean intersection over Union score of 0.95 or better for each region in each category.
Time frame: 6 months
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.