Correlation of Image Quality Metrics with Expert Perception for Industrial Computed Tomography

This research investigates the effectiveness of image quality metrics in predicting human perception of superresolution computed tomography images. By comparing multiple metrics and correlating their scores with human expert ratings on image attributes such as noise, contrast, and sharpness, we aim to identify metrics that closely align with human visual judgment. Our findings highlight the limitations of relying solely on a single metric for image quality assessment and emphasize the importance of considering specific tasks and image content when selecting appropriate metrics. This study contributes to a deeper understanding of the relationship between objective image quality metrics and subjective human perception in the context of industrial computed tomography.