SynQ-ViT: Synthetic Image Quality Assessment for CT Calibration with Vision Transformer

Optimizing the Diagnostic Imaging System (DIS) of Image Quality Assessment (IQA) in Computed Tomography (CT) imaging using contrast media requires considerate evaluation. In this paper, we present a novel methodology, called SynQ-ViT, for evaluating CT image quality. SynQ-ViT involves restructuring the model to optimize memory usage and processing speed. The dataset comprises CT images of a Perspex phantom, injected with gold nanoparticles as contrast media, and collected with varying exposure settings. SynQ-ViT, achieved an R-squared value of 0.8892, outperforming other state-of-the-art models with fewer parameters, demonstrating its efficiency in predicting Signal-to-Noise Ratio (SNR) and Contrast-to-Noise Ratio (CNR) values that can be used for optimal calibration of CT scanners automatically or semi-automatically, leveraging image quality and hence medical diagnostics. We anticipate this approach can maintain image fidelity, particularly in low-resolution contexts.