Denoising of Nifti (MRI) Images with a Regularized Neighborhood Pixel Similarity Wavelet Algorithm

Area of Science:

• Medical Imaging
• Image Processing
• Signal Processing

Background:

• Image corruption from noise poses a significant challenge in image processing, obscuring features and biasing analysis.
• Accurate noise reduction is crucial for reliable interpretation of medical imaging data, such as magnetic resonance imaging (MRI).

Purpose of the Study:

• To develop and evaluate a novel algorithm for denoising Nifti MRI images.
• To improve the accuracy and efficiency of noise reduction in corrupted images.

Main Methods:

• The Regularized Neighborhood Pixel Similarity Wavelet (PixSimWave) algorithm was developed.
• It utilizes regularized pixel similarity detection and adaptive neighborhood filtering for noise reduction.
• Wavelet transform is employed for sparse image representation and soft thresholding.

Main Results:

• PixSimWave achieved high Structural Similarity Index (SSIM) values (0.9908 at 3% Rician noise, 0.9881 at 17% Rician noise).
• The method demonstrated superior performance compared to other denoising techniques on simulated and raw T1w MRIs.
• Improved Peak Signal-to-Noise Ratio (PSNR) and SSIM were observed with the addition of Gaussian noise, preserving edges and textures.

Conclusions:

• The PixSimWave algorithm is a viable and effective approach for noise elimination in MRI.
• It combines sparse wavelet coefficients and regularized similarity for accurate noise reduction with decreased computation time.
• The method shows significant potential for enhancing the quality of medical imaging data.