Seasons Of Code

Automated Fiducial Localisation from MRI/CT Images    • Ritwick Chaudhry    • Shubham Goel   

Automated Fiducial Localisation from MRI/CT Images

Automated Fiducial Localisation from MRI/CT Images


This project aims to automate the marking of the fiducials (bone based markers) in MRI/CT scan images.

Problem Statement:

Image-guided neurosurgery systems (IGNS) play an important role in intracranial surgery. These systems have a live CT/MRI scanning during surgeries to aid the doctor. During neuro-registration, bone based markers or fiducials are affixed to the skull before imaging so that they can calibrate the coordinate systems of the imaging device and the actual 3D World Coordinates. As of now, the actual Fiducials are tapped by instruments, and on the other hand they are MANUALLY marked out in the images obtained from the MRI/CT imaging. Automatic localization of these centers will help to reduce the human error in registration and speed up the registration process. In a nutshell, the task is to automatically locate these Fiducial markers from these images, so that they don’t have to be manually marked out by the doctor.

Practical Details:

This project was a research challenge provided by scientists working in BARC (Bhabha Atomic Research Centre), during the Inter IIT Tech Meet, 2017 held at IITM. We came second, after developing an algorithm which worked quite well. We believe it can be further worked upon to work extremely well. This task is still open, and scientists in India are actively looking for a solution. Also, we found very little literature on this task, and aim to publish this work too.