MRI intensity inhomogeneity correction by combining intensity and spatial information.

Area of Science:

• Medical Imaging
• Image Analysis
• Computational Neuroscience

Background:

• Intensity inhomogeneity is a common artifact in medical imaging, particularly MRI.
• This artifact significantly impacts the accuracy of automated image analysis and quantitative assessments.
• Existing correction methods may lack efficiency or adaptability to dynamic variations.

Purpose of the Study:

• To develop a fully automated, non-parametric method for retrospective correction of intensity inhomogeneity.
• To improve the dynamic correction of local intensity variations using spatial image features.
• To enhance the reliability of quantitative analysis in medical imaging tasks.

Main Methods:

• A four-step iterative procedure incorporating intensity and spatial image features.
• Computation of intensity correction forces based on probability distributions and second derivatives.
• Regularization of voxel forces to estimate the inhomogeneity correction field.
• Dynamic correction adjustment controlled by the regularization kernel size.

Main Results:

• The proposed method effectively corrects dynamic intensity inhomogeneity in simulated and real MR brain images.
• Qualitative and quantitative evaluations confirm successful artifact removal.
• The method demonstrated robustness, not corrupting inhomogeneity-free images.

Conclusions:

• The novel automated method provides efficient and dynamic correction of intensity inhomogeneity.
• This technique improves the foundation for accurate quantitative analysis in medical imaging.
• The approach offers a significant advancement in retrospective artifact correction for MRI data.