A customized acutance metric for quality control applications in MRI

Area of Science:

• Medical Imaging
• Image Processing
• Computational Science

Background:

• Acutance, a subjective measure of image edge quality, is crucial for optimizing magnetic resonance imaging (MRI) protocols and post-processing algorithms.
• Existing objective metrics for image blur and sharpness are primarily designed for natural scenes and are not suitable for MR imagery.
• There is a need for a specialized, efficient metric to quantify acutance in MRI applications.

Purpose of the Study:

• To introduce a novel, computationally fast objective metric for measuring image sharpness and blur specifically tailored for MRI.
• To address the lack of suitable acutance metrics in the field of magnetic resonance imaging.

Main Methods:

• A new metric, the largest local gradient-based sharpness metric (LLGSM), was developed.
• The LLGSM calculates the root mean square (RMS) of exponentially weighted largest local gradient (LLG) values, ordered lexicographically and sorted in descending order.
• The metric's performance was evaluated against subjective acutance scores and computational efficiency.

Main Results:

• The proposed LLGSM demonstrates high efficiency in measuring sharpness and blur in MR imagery.
• LLGSM shows superior overall agreement with subjective acutance assessments compared to existing alternative metrics.
• The metric offers significant computational speed advantages.

Conclusions:

• The LLGSM is an effective and efficient objective metric for evaluating image acutance in MRI.
• This metric can aid in the design and optimization of MRI imaging protocols and image processing techniques.
• LLGSM provides a valuable tool for enhancing the quality of MR imagery.