Improving the performance of the prony method using a wavelet domain filter for MRI denoising

Area of Science:

• Medical Imaging
• Signal Processing
• Biomedical Engineering

Background:

• Magnetic Resonance Imaging (MRI) is crucial for diagnosing brain abnormalities.
• Rician noise is a common artifact in T2-weighted MRI, potentially hindering accurate tissue classification.
• Existing noise reduction methods may not sufficiently preserve image details for precise analysis.

Purpose of the Study:

• To develop and evaluate a novel wavelet domain bilateral filtering technique for Rician noise reduction in T2-weighted MRI.
• To assess the effectiveness of the proposed filter in improving abnormal brain tissue detection.
• To compare the performance of the new filter against existing methods.

Main Methods:

• A modified wavelet domain bilateral filter based on Kazubek's algorithm was developed.
• The filter was applied to synthetic and real T2-weighted MRI sequences.
• A variant of the Prony method was employed for exponential fitting and tissue classification.
• Quantitative and qualitative criteria were used for performance evaluation.

Main Results:

• The new wavelet domain bilateral filter demonstrated significant Rician noise suppression capabilities.
• The filter effectively preserved important image features while reducing noise.
• Abnormal brain tissue detection was enhanced following the application of the proposed filtering technique.

Conclusions:

• The developed wavelet domain bilateral filter is a promising tool for noise reduction in T2-weighted MRI.
• This method improves the accuracy of abnormal brain tissue detection, aiding in clinical diagnosis.
• Further validation on diverse datasets is warranted to confirm its clinical utility.