Image Quality Metrics and Radiological Evaluation
Author Information
Author(s): Marchetto Elisa, Eichhorn Hannah, Gallichan Daniel, Schnabel Julia A., Ganz Melanie
Primary Institution: Cornell University
Hypothesis
The study aims to compare the performance of reference-based and reference-free image quality metrics in the presence of motion artifacts.
Conclusion
Reference-based metrics reliably correlate with radiological evaluation across different sequences and datasets.
Supporting Evidence
- All reference-based image quality metrics showed strong correlation with observer assessments.
- Average Edge Strength among reference-free metrics displayed stronger correlations across all sequences.
- Normalization and brain masking significantly influence correlation values.
Takeaway
This study looked at how well different ways of measuring image quality match up with doctors' evaluations, especially when there are motion problems in the images.
Methodology
The study compared five reference-based and five reference-free image quality metrics on datasets with and without motion, using a Likert scale for ratings and calculating Spearman correlation coefficients.
Limitations
The study may be limited by the specific datasets used and the subjective nature of radiologist assessments.
Want to read the original?
Access the complete publication on the publisher's website