Biased motion-adaptive temporal filtering for speckle reduction in echocardiography

Area of Science:

• Medical Imaging
• Signal Processing
• Biomedical Engineering

Background:

• Speckle noise significantly degrades ultrasound image quality.
• Automated echocardiographic interpretation requires effective noise reduction techniques.
• Existing filtering methods can cause blurring or fail to preserve important image features.

Purpose of the Study:

• To develop a novel motion-adaptive spatio-temporal filtering technique for ultrasound images.
• To reduce speckle noise while preserving critical image structures, particularly boundaries.
• To improve the input quality for subsequent automated echocardiographic image analysis.

Main Methods:

• Implementation of a motion-compensated Two-Dimensional Least Mean Squares (TDLMS) filter.
• Utilizing the mode of the intensity distribution for a biased temporal filter output.
• Development of an adaptive algorithm for filter convergence control.
• Evaluation using simulated, phantom, and in vivo carotid artery ultrasound sequences.

Main Results:

• The Biased TDLMS filter effectively reduces speckle noise in ultrasound images.
• Quantitative analysis demonstrates superior edge preservation compared to other techniques.
• Filtered images show improved contrast and reduced noise, beneficial for feature extraction.
• Successful application in echocardiographic boundary detection, including left atrium and mitral valve analysis.

Conclusions:

• The proposed motion-adaptive spatio-temporal filter significantly enhances ultrasound image quality.
• The Biased TDLMS filter offers a robust solution for noise reduction and edge preservation in echocardiography.
• This technique provides better image inputs for automated echocardiographic interpretation schemes.