Multi-Energy CT Material Property Extraction via Statistical Analysis

The patent applies parameter changes by utilizing multiple energy levels (kVp values) for CT scanning. Instead of using a single energy level, the system acquires images at different energies (e.g., 80kVp, 140kVp) and processes these multi-energy datasets to differentiate materials. This allows the system to distinguish between materials like iodine and calcium that appear similar in single-energy images, thereby improving material differentiation accuracy without requiring fundamentally new imaging hardware

The patent introduces an additional dimension by adding energy level variation to the traditional CT imaging process. By acquiring images across multiple energy levels and using statistical analysis (e.g., histograms, machine learning algorithms) on this extended data space, the system creates a more robust basis for material identification. This dimensional expansion allows differentiation based on energy-dependent attenuation characteristics that are not visible in conventional single-energy images

The patent implements preliminary action by performing statistical analysis and material classification algorithms during the image reconstruction phase, rather than requiring separate measurement steps. The system pre-processes the multi-energy CT data using histograms and machine learning models to generate material property maps simultaneously with image formation, thereby improving measurement precision without adding significant time to the overall workflow

The patent uses calibration curves and reference material databases that store pre-measured attenuation characteristics for various materials at different energy levels. During analysis, the system compares patient data against these pre-established references to rapidly determine material properties, improving measurement accuracy while minimizing additional scanning or processing time required