Related Experiment Video
Updated: Jul 16, 2026
Simultaneous Data Collection of fMRI and fNIRS Measurements Using a Whole-Head Optode Array and Short-Distance Channels
Published on: October 20, 2023
Many heads are better than one: jointly removing bias from multiple MRIs using nonparametric maximum likelihood
Erik G Learned-Miller1, Vidit Jain
1Department of Computer Science, University of Massachusetts, Amherst, MA 01003, USA. elm@cs.umass.edu
Abstract:
The correction of multiplicative bias in magnetic resonance images is an important problem in medical image processing, especially as a preprocessing step for quantitative measurements and other numerical procedures. Most previous approaches have used a maximum likelihood method to increase the probability of the pixels in a single image by adaptively estimating a correction to the unknown image bias field. The pixel probabilities are defined either in terms of a pre-existing tissue model, or nonparametrically in terms of the image's own pixel values. In both cases, the specific location of a pixel in the image does not influence the probability calculation. Our approach, similar to methods of joint registration, simultaneously eliminates the bias from a set of images of the same anatomy, but from different patients. We use the statistics from the same location across different patients' images, rather than within an image, to eliminate bias fields from all of the images simultaneously. Evaluating the likelihood of a particular voxel in one patient's scan with respect to voxels in the same location in a set of other patients' scans disambiguates effects that might be due to either bias fields or anatomy. We present a variety of "two-dimensional" experimental results (working with one image from each patient) showing how our method overcomes serious problems experienced by other methods. We also present preliminary results on full three-dimensional volume correction across patients.
Insights
This study introduces a novel method for correcting magnetic resonance imaging bias fields by analyzing voxel statistics across multiple patients. This approach improves quantitative measurements by simultaneously correcting bias in a set of images.
Area of Science:
- Medical Image Processing
- Computational Biology
- Radiology
Background:
- Multiplicative bias in magnetic resonance (MR) images hinders quantitative analysis.
- Existing methods often estimate bias fields within a single image, ignoring spatial location.
- Previous techniques rely on tissue models or nonparametric pixel value distributions.
Purpose of the Study:
- To develop a novel method for correcting multiplicative bias in MR images.
- To improve the accuracy of quantitative measurements in medical imaging.
- To address limitations of single-image bias correction methods.
Main Methods:
- A novel approach that simultaneously corrects bias fields across a set of MR images from different patients.
- Utilizes statistics from the same anatomical location across multiple patient scans.
- Evaluates voxel likelihoods across patient cohorts to distinguish bias from anatomy.
Main Results:
- Demonstrates superior performance over existing methods in correcting bias fields.
- Presents successful "two-dimensional" experimental results using single images per patient.
- Shows preliminary success in three-dimensional volume correction across patients.
Conclusions:
- The proposed cross-patient bias correction method effectively addresses limitations of single-image approaches.
- This technique enhances the reliability of quantitative MR imaging.
- The method shows promise for both 2D and 3D MR image bias correction.
