An Iterative Method for Computed Tomography Machine Calibration: Attenuation and Dual-Energy Computed Tomography Inversion from the Hounsfield Numbers

Photoelectric effect and X-ray scattering determine the attenuation coefficient of materials in diagnostic radiology. This manuscript presents an iterative gradient search method to separate the contributions to attenuation from these two independent sources. This issue assumes importance due to two reasons, including 1) Electron density determination of scanned materials and 2) correct dose calculation in diagnostic radiology. A special water-filled phantom which was custom-built for simultaneous scanning of 12 samples was used in the current study. Attenuation coefficient equations were iteratively solved to calculate the contributions from x-ray scattering and photoelectric effects. Data converged after five iterations (within 1%). Error in the attenuation coefficient was measured at ±3%. As evidenced by the obtained results, this method can be used to determine the Compton and photoelectric contributions with sufficient accuracy. Moreover, the inversion of Dual- Energy computed tomography (DECT) data for finding electron density and effective atomic number of materials also presents satisfactory results.