Edge-preserving speckle texture removal by interference-based speckle filtering followed by anisotropic diffusion

Area of Science:

• Medical Imaging
• Signal Processing
• Biomedical Engineering

Background:

• Speckle noise is an inherent artifact in ultrasonography, degrading image quality and hindering accurate object recognition.
• This noise arises from the constructive and destructive interference of scattered wavefronts, leading to image texture and reduced diagnostic information.

Purpose of the Study:

• To develop and evaluate a novel filter, the interference-based speckle filter followed by anisotropic diffusion (ISFAD), for speckle noise reduction in B-mode ultrasound images.
• To assess the filter's ability to preserve image edges and local gray-level information while removing speckle texture.

Main Methods:

• The proposed ISFAD filter was designed to attenuate constructive and destructive interference patterns within ultrasound images.
• Performance evaluation involved comparing ISFAD against 10 other filters using simulated ultrasound images from Field II.
• Quantitative assessment included structural similarity index and analysis of segmentation task performance (true positive, false positive, accuracy).

Main Results:

• The ISFAD filter achieved the highest structural similarity score (0.95) when compared to 10 other filters.
• Functional improvement in image segmentation was demonstrated, with ISFAD yielding accuracy rates exceeding 90% for structures with well-defined borders across three segmentation techniques.

Conclusions:

• The ISFAD filter effectively reduces speckle noise in B-mode ultrasound images while preserving crucial image details like edges and gray levels.
• ISFAD significantly enhances the accuracy of image segmentation, offering a valuable tool for improving diagnostic capabilities in ultrasonography.