Image nonuniformity correction in high field (1.5T) MRI

Feroze B. Mohamed, Simon Vinitski, Carlos Gonzalez, Scott Faro, Claudio Burnett, Hector V. Ortega, Tad Iwanaga

Research output: Contribution to journalConference articlepeer-review


The largest source of image nonuniformity in high field MRI systems arises from the specialized radiofrequency (RF) coils that are currently used. It degrades conspicuity of lesion(s), and reduces accuracy of image post-processing (e.g. MRA, tissue segmentation, etc.). In this investigation we devised a method to correct nonuniformity of MR images with correction matrices obtained from cylindrical uniform phantom. The phantom, filled with doped water and exactly fitted the volume of the head coil, was imaged using MRI/MRA clinical pulse sequences. The reference pixel intensity was defined using the most homogeneous region of the RF coil, and every voxel inside the phantom was normalized relative to the reference value. A correction matrix was obtained for each type of MRI contrast. MRI and MRA images of phantoms as well as normals and MS patients were obtained and processed by the correction matrices. Application of the correction matrices to phantom data resulted in up to 20 fold improvement in image uniformity. In humans, the corrected images improved sharpness and tissue contrast, leading to increased conspicuity of the lesions.

Original languageEnglish (US)
Pages (from-to)471-472
Number of pages2
JournalAnnual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings
Issue number1
StatePublished - Dec 1 1995
EventProceedings of the 1995 IEEE Engineering in Medicine and Biology 17th Annual Conference and 21st Canadian Medical and Biological Engineering Conference. Part 2 (of 2) - Montreal, Can
Duration: Sep 20 1995Sep 23 1995

ASJC Scopus subject areas

  • Signal Processing
  • Biomedical Engineering
  • Computer Vision and Pattern Recognition
  • Health Informatics


Dive into the research topics of 'Image nonuniformity correction in high field (1.5T) MRI'. Together they form a unique fingerprint.

Cite this