PyJointTrack has a GUI!

Post date: Sep 07, 2009 5:34:17 AM

The final phase of my data analysis will involve more image registration, this time between images that i collected from cadaver spines in our lab. One of the biggest difference between this phase and phase 2 is that i will have to generate my own starting poses, and will not simply be able to script them from a file. Another big difference is that i will be analyzing image sequences instead of static images. Therefore, i will need a better way to manipulate and process my data that the hardcoded method that i have been using so far. I have been working hard this weekend to implement a simple GUI for my registration tool and i finally have something to show for it. It only does the most basic things right now like load images (almost any format), volumes and calibration data and will perform a registration using the default position of the volume as a starting pose. One particularly cool feature is that it can save and load project files (image file paths, calibration data, poses, centers of rotation, regions of interest, optimizer settings, metric settings, registration settings, patient/trial info). The project data is saved in an xml file and can be shared with other applications (like Shangs JointTrack). Also, since the file is saved as text it is possible (though not necessarily advisable) to edit the file by hand. Next steps:

    1. Fix user interaction for DRR, rotations not working

    2. Assign the settings from the project file to the correct variables in the application

    3. Figure out how to handle multiple regions of interest (vertebral bodies) on one volume

    4. Determine what the best way to get good starting positions (image overlay seems to cluttered, outlines may not give enough detail)