TECH OFFER
A Transfer Learning Assisted 3D Deep Learning Method For Knee Osteoarthritis Detection
TECHNOLOGY OVERVIEW
The invention introduces a method for detecting and determining knee osteoarthritis (OA) from medical imaging, specifically using a transfer learning assisted 3D deep learning approach. The process involves obtaining magnetic resonance imaging (MRI) volumes, converting them to the Neuroimaging Informatics Technology Initiative (NIfTI) standard, and transforming them to a standardized format through resizing, Z-normalizing, and standardizing. Feature extraction is then performed using different neural networks. The method trains and compares MRI volumes using both pretrained and randomly assigned initial weights, utilizing pretrained weights as the initial weights for the training process of 3D models. The performance of these models is assessed based on classification metrics to effectively detect knee osteoarthritis.
Mega - Trends
Wireless Intelligence, 4 IR, Innovating to Zero, Future Broadband Applications, Generation Y, SMART Clouds: The Nest Trend in Cloud Computing, Future Fortune Global 500, Factory of the Future: SMART and GREEN, SMART Cities and SMART Infrastructure
Technology Readiness Level (TRL)
TRL 3
Patent Number
PI 2023002787
Get the technology fact sheet here:
Contact person for this offer:
ChM Dr. Lee Ching Shya, PhD (Dual), RTTP
Technology Transfer Manager
Email: leecs@um.edu.my
Tel: +603-7967-7351/ 013-2250151