Low-Dose X-Ray Preview via Correlated Noise Injection

The patent applies local quality by distinguishing between different detector types and applying appropriate noise models to each. For flat-panel detectors, correlated noise models are used specifically at the detector level, while uncorrelated noise assumptions can be used for multi-detector CT scanners. This localized adaptation allows each detector type to have its specific noise characteristics accurately reproduced without compromising the versatility of the overall dose reduction simulation framework.

The patent applies preliminary action by performing low-dose simulation before the actual CBCT scan. The simulation uses correlated noise models to predict the image quality that would result from various dose levels, allowing clinicians to select the minimum adequate dose in advance. This preliminary assessment prevents both over-dosing (excessive radiation) and under-dosing (unacceptable image quality) by providing accurate noise predictions at different dose levels before the actual scan occurs.

The patent applies parameter changes by enabling dynamic adjustment of dose levels for each repeat CBCT scan based on the specific imaging task requirements. The correlated noise simulation allows clinicians to evaluate different dose parameters for different imaging tasks (e.g., setup vs. verification scans) and select the minimum adequate dose for each. This parameter optimization across multiple scans significantly reduces the total accumulated radiation dose while ensuring each scan meets its specific imaging task requirements.