Motion Monitor for movement in MRI using reflective marker, optical fibers and webcam

Open Access
Wander, Abhijit Singh
Graduate Program:
Electrical Engineering
Master of Science
Document Type:
Master Thesis
Date of Defense:
October 09, 2014
Committee Members:
  • Robert Anthony Gray, Thesis Advisor
  • MRI
  • motion detection
  • optical fibers
  • webcam
Magnetic resonance imaging (MRI) is an imaging technique used as a research modality for the study of human anatomy. Patient motion during MRI scanning remains a severe problem which degrades the diagnostic accuracy, posing significant problems in the acquisition and analysis of the MRI data. In this thesis, three algorithms were explored for real time motion detection of head movement in patients. The current motion detection methods are expensive and, to a degree, scanner specific making it problematic for inclusion in every scanner. The proposed research is a potential alternative, as it is scanner independent and based on low cost hardware namely webcam, non-coherent optical fiber bundles, reflective marker and a personal computer. One end of the optical fiber bundle is above the reflective marker placed on the head of patient and the other end is connected to the webcam. The proposed algorithms use the red, green and blue channels of the color of light emerging from the output end of the optical fiber bundle to determine the range of movement. Bench testing demonstrated that two of the algorithms seemed to have some shortcomings to be usable at this time. The third algorithm passed the bench tests outside the scanner for head-foot linear movement (like straightening of a body when lying on MRI table) and roll rotational movement (like shaking of the head for a no). Head-foot linear movement and pitch rotational movement (like nodding of the head for a yes) were varied to some degrees during calibration inside the scanner. To evaluate the effects of the movements on the resulting MRI images, MRI scan was taken to acquire images with motion. A simulation analysis was performed on an MRI image dataset obtained from internet. The results of the simulation were compared with the results of the images with motion acquired in the MRI scan. The tests demonstrated that motion monitor was able to detect movements as small as 2.5 mm and 1 degree for the head-foot and roll movements respectively. A feedback system can be developed to inform the patients of movement which will help the patients to remain still during the MRI scan. The merits of the motion monitor device are its real time capability, scanner independent, low cost hardware and patient comfort.