Efficient segmentation and correction model for brain MR images with level set framework based on basis functions.

Area of Science:

• Medical Imaging
• Image Processing
• Computational Biology

Background:

• Magnetic Resonance (MR) imaging is crucial for brain disease detection.
• MR images suffer from intensity inhomogeneity (bias field), hindering accurate analysis.
• Existing bias field correction methods often result in poor segmentation or over-correction.

Purpose of the Study:

• To develop a new model for simultaneous segmentation and moderate bias field correction in MR images.
• To address limitations of current methods, particularly over-correction issues.
• To enhance the accuracy and reliability of MR image analysis.

Main Methods:

• The proposed model is based on the Multiplicative Intrinsic Component Optimization (MICO) model.
• It integrates the split Bregman method for simultaneous segmentation and bias correction.
• The model was applied and validated on a large dataset of MR images.

Main Results:

• The new model effectively segments tissues and corrects bias fields in MR images.
• Experimental results demonstrate superior performance compared to the MICO model.
• The model shows accuracy, efficiency, and robustness in handling intensity inhomogeneity.

Conclusions:

• The developed model offers a significant improvement for MR image segmentation and bias field correction.
• It provides a more moderate and simultaneous approach, overcoming limitations of prior methods.
• The findings suggest enhanced potential for accurate disease detection using MR imaging.