Image registration is a crucial and fundamental procedure in Image Guided Radiation Therapy (IGRT). We aim to develop a deep learning-based algorithm for deformable medical image registration for IGRT. We would like to take advantage of large scale public datasets from different organizations to train and test the presented deep learning-based deformable registration algorithm. Moreover, we wish to develop a robust deformable registration algorithm that can perform well on generic data. The demographic data from the report could be used to obtain more information about the population distribution, which could be used in our further analysis.

1. Develop a learning-based algorithm for deformable medical image registration.
2. Using different datasets to accomplish cross-validation on deep learning with the help of demographic data.

Zhihe Zhao, Shenzhen institutes of advanced technology, Chinese Academy of Sciences.
Jiarui Zhu, Shenzhen institutes of advanced technology, Chinese Academy of Sciences.