FocalUNETR

A Focal Transformer for Boundary-aware Prostate Segmentation using CT Images (MICCAI20223, paper link)

Data Preprocessing

• You can refer to this link for generating boundaries from the ground truth mask of the organ you would like to segment.
• Please follow this reference regarding your original datasets for conversion, such as operations for resampling, cropping, and padding.

For Training and Testing

• Based on the settings in the option folder containing yml files, training and testing can be performed with or without using the boundary-aware regression auxiliary task.
• For additional baseline models, please refer to this link.

TO DO

• Tutorials for inference and visualization.
• ..

Contact

• Chengyin Li, cyli@wayne.edu