[Curvelet denoising algorithm for medical ultrasound image based on adaptive threshold]

Area of Science:

• Medical Imaging
• Signal Processing
• Biomedical Engineering

Background:

• Traditional ultrasound image denoising algorithms often remove essential details and edge information during speckle noise suppression.
• Speckle noise significantly degrades the quality and diagnostic accuracy of ultrasound images.

Purpose of the Study:

• To propose a novel adaptive thresholding algorithm based on curvelet transform for enhanced ultrasound image denoising.
• To effectively reduce speckle noise while preserving fine details and edge information in ultrasound images.

Main Methods:

• The proposed algorithm utilizes the curvelet transform to decompose ultrasound images.
• It analyzes local variance differences in coefficients across image layers to define fuzzy regions and membership functions.
• An adaptive threshold, determined by these membership functions, is applied for denoising.

Main Results:

• The algorithm demonstrates effective reduction of speckle noise in ultrasound images.
• It successfully retains important detail and weak edge information, outperforming traditional methods.
• Experimental results indicate a significant enhancement in the performance of B-mode ultrasound instruments.

Conclusions:

• The adaptive thresholding algorithm based on curvelet transform offers superior speckle noise reduction for ultrasound images.
• Preservation of image details and edges leads to improved diagnostic quality.
• This method has the potential to enhance the overall performance and utility of ultrasound imaging devices.