Primary Contributor: Rohini Prabhakar Gaikar
The limited number of manually labelled data is the biggest challenge in implementation supervised training algorithms in medical image processing. Therefore, we have implemented the transfer learning approach to improve fully automated kidney segmentation on small multi-parametric (mp-MRI) dataset sequences using Attention U-Net deep learning model. Due to the difference in acquisition protocols in mp-MRI sequences a single model will not generalize on all mp-MRI sequence images, but the knowledge gained by model during T1-Weighted contrast enhanced Nephrographic Phase (T1W-NG) training when transferred to small size target dataset (T2W, T1-CM, T1-PRE, T1-IP, and T1-OOP) improved the kidney segmentation results.
Use the following links for more information:
http://dx.doi.org/10.1117/1.JMI.9.3.036001
https://doi.org/10.1117/12.2607526