Using Machine Learning Software (Deep Anatomical Federated Network) to Analyze Magnetic Resonance Imaging (MRI) Sequences of Anatomical Structures

We aim to enhance AI learning of MR images by providing scans of the pectoralis major, deltoid, and gluteus maximus from young healthy adults to Deep Anatomical Federated Network (DAFNE). DAFNE is a decentralized AI software that improves over time through user refinement. Machine Learning (ML) models enable faster automated analysis, distinguishing anatomical structures and abnormalities.
Currently, DAFNE has not analyzed the muscle groups our research will provide. Training ML models for MR scans offers Built With Science (BWS) an effective tool for segmenting muscle boundaries, helping assess the impact of training interventions on muscle size and shape. This supports BWS in communicating the latest evidence on improving body composition.
By contributing images and refining models, we enhance DAFNE’s efficiency for future users. BWS can continue supplying MR scans, improving DAFNE while benefiting from its evolving capabilities. Our MR scans will be conducted at the UBC MRI Research Facility, University of British Columbia, Vancouver, Canada.

Faculty Supervisor:

Cameron Mitchell

Student:

Partner:

J. Ethier Holdings Corp

Discipline:

Life Sciences

Sector:

Education

University:

The University of British Columbia

Program: