Publications by authors named "Tu Ren-Zhe"
Article Synopsis
- Deep brain stimulation (STN-DBS) is a key treatment for advanced Parkinson's disease, where the success of surgery depends on precisely placing the electrode in the subthalamic nucleus.
- This study proposes an automated deep learning approach to improve target localization for STN and red nucleus segmentation, addressing challenges such as ambiguous boundaries and variable target shapes.
- The developed model integrates a hierarchical attention mechanism within an encoder-decoder framework and was tested on MRI data from 99 Parkinson's patients, achieving strong segmentation metrics like an 88.20% Dice similarity coefficient.
