Edge Sharpness Assessment by Parametric Modeling: Application to Magnetic Resonance Imaging.

Area of Science:

• Biomedical Imaging
• Medical Image Analysis
• Image Quality Assessment

Background:

• Edge sharpness is a critical, yet often underestimated, metric for evaluating biomedical image quality.
• Assessing edge sharpness in noisy images, such as those from magnetic resonance imaging (MRI), presents significant challenges.

Purpose of the Study:

• To develop and validate a semi-automatic method for quantifying edge sharpness in noisy biomedical images.
• To apply this method to magnetic resonance imaging (MRI) data, demonstrating its utility in challenging imaging conditions.

Main Methods:

• A semi-automatic approach using parametric modeling of image edges.
• Generation of an edge map, manual selection of edges-of-interest (EOI), and application of exclusion criteria.
• Fitting intensity profiles normal to the EOI with a sigmoid function to extract sharpness parameters.

Main Results:

• The method successfully quantifies edge sharpness in simulated and real MRI data (phantom and cine imaging).
• Demonstrated effectiveness in images with poor signal-to-noise ratio.
• The distribution of the sharpness parameter provides a quantitative measure of edge sharpness.

Conclusions:

• The developed method offers a fast and quantitative evaluation of edge sharpness in noisy biomedical images.
• The technique is particularly valuable for magnetic resonance imaging (MRI) but is adaptable to other medical imaging modalities.