Constructing Multiwavelet-based Shearlets and using Them for Automatic Segmentation of Noisy Brain Images Affected by...

Area of Science:

• Medical Imaging
• Signal Processing
• Computational Neuroscience

Background:

• COVID-19, caused by SARS-CoV-2, primarily affects the respiratory system but recent findings indicate potential neurological impact.
• The brain's central role in bodily functions necessitates investigating its health in the context of COVID-19.
• Brain image segmentation aids in identifying coronavirus-affected regions, but image noise in MRA scans presents a challenge.

Purpose of the Study:

• To introduce and evaluate multishearlets constructed from multiwavelets for denoising brain MRA images.
• To develop an automatic scheme for initial curve characterization in active contour models for improved segmentation.
• To assess the efficacy of multishearlets in detecting COVID-19-related brain abnormalities.

Main Methods:

• Construction of multishearlets based on multiwavelets, expanding directional properties.
• Application of multishearlets for denoising brain magnetic resonance angiography (MRA) images.
• Development of an automatic active contour model segmentation scheme with initial curve detection.

Main Results:

• Multishearlets demonstrated superior denoising performance compared to original shearlets, evidenced by higher Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM).
• The proposed segmentation scheme achieved an accuracy of 0.99.
• Multishearlets effectively neutralized noise in MRA images.

Conclusions:

• Multishearlets offer a significant improvement over traditional shearlets for denoising MRA images.
• The developed segmentation method provides highly accurate results for identifying affected brain regions.
• This research contributes a robust method for analyzing neurological effects of COVID-19 using enhanced image processing techniques.