A New Adaptive Diffusive Function for Magnetic Resonance Imaging Denoising Based on Pixel Similarity.

Area of Science:

• Medical Image Processing
• Computational Imaging
• Signal Processing

Background:

• Partial differential equation (PDE)-based denoising is crucial for medical imaging like MRI.
• Existing Perona-Malik (P-M) models struggle with low signal-to-noise ratio (SNR) images due to reliance solely on gradient information.

Purpose of the Study:

• To propose a modified diffusive function with fractional power, enhancing the P-M model for low SNR medical images.
• To demonstrate the stabilization of the P-M method using the proposed function.
• To improve noise removal and edge preservation in low SNR environments.

Main Methods:

• Development of a modified diffusive function incorporating pixel similarity and fractional power.
• Application of the modified function within a PDE-based denoising framework.
• Experimental validation comparing the proposed method against standard P-M functions.

Main Results:

• The proposed fractional power diffusive function significantly improves SNR in low SNR images.
• Enhanced edge preservation capabilities compared to traditional P-M functions.
• Demonstrated stabilization of the P-M denoising method.

Conclusions:

• The modified diffusive function offers a superior approach for denoising low SNR medical images.
• This method effectively balances noise reduction with the preservation of critical image features.
• The proposed technique advances PDE-based denoising for medical applications.