Automatic motion correction of Musculoskeletal MRI using DSLR camera.

Autor: Chikop SA; Medical Imaging Research Centre, Dayananda Sagar Institutions, India., Anchan ABS; Medical Imaging Research Centre, Dayananda Sagar Institutions, India., Koulagi G; Medical Imaging Research Centre, Dayananda Sagar Institutions, India., Honnedevasthana AA; Medical Imaging Research Centre, Dayananda Sagar Institutions, India., Imam S; Medical Imaging Research Centre, Dayananda Sagar Institutions, India., Geethanath S; Medical Imaging Research Centre, Dayananda Sagar Institutions, India; Magnetic Resonance Research Program, Columbia University in the City of New York, New York, USA. Electronic address: sg3606@columbia.edu.
Jazyk: angličtina
Zdroj: Magnetic resonance imaging [Magn Reson Imaging] 2018 May; Vol. 48, pp. 74-79. Date of Electronic Publication: 2018 Jan 05.
DOI: 10.1016/j.mri.2017.12.031
Abstrakt: The purpose of this study is to illustrate motion correction in Musculoskeletal (MSK) Magnetic Resonance Imaging (MRI) through utilization of information from an optical tracker to capture the extent and instant of motion. A Digital Single Lens Reflexive camera is employed as the optical tracker to capture the extent and instant of motion. A checkerboard is utilized as a marker that is placed on the coil. Shift of the checkerboard provides the extent of motion, which is captured by camera and is used for motion correction in (MSK)-MRI images. Experiments were first performed on an in vitro phantom to obtain calibration curves, which determine the relationship between object movement and pixel shifts. Six healthy volunteers were recruited for the study and experiments were repeated thrice on each subject. Reducing the gradient entropy of the image with reference to the calibration curve resulted in motion correction. Fusion of motion-free data with motion-corrupted data and motion free data with motion-corrected data was performed for qualitative analysis of data. Normalized Root Mean Squared Error of the motion-corrected data with respect motion-free was approximately 20% lesser compared to motion-corrupted data with respect to motion-free data with better delineation of edges and reduced ghosting. The work focuses on time of displacement through an external tracker on the RF coil and utilizes that information for motion correction. The method can be readily implemented on a clinical scanner, while it is not necessary for the subject to wear motion sensors.
(Copyright © 2018 Elsevier Inc. All rights reserved.)
Databáze: MEDLINE
Externí odkaz: