Approved Research
Deep learning based multi-modal medical image registration and segmentation
Lay summary
- using the multimodality brain imaging, we aim to develop novel multimodality registration algorithms using deep learning, which extract features from images and optimize the parameters.
The developed method could be used to perform planning before a brain resection surgery,
and later used for image navigation during a brain surgery.
- to better understand brain functions, using brain parcellation which labeling the areas of brain and multi-atlas methods, perform brain function analysis
- compare brain imaging with cognitive functions
- the proposed project timeline would be 2-3 years, and hopefull a research-level pre-surgical imaging planning algorithm and software are developed in this project
The impact to public health:
In an aging society, this study could help to better understand brain development and cognitive functions, which could help to alleviate aging problems, and hopefully solve brain degenerative diseases such as Alzheimers' , Parkinson's diease, Epilepsy or depression
It would also help to predict brain disease outcomes before surgery, as well as to better evaluate the patient's progression over a long period.