Performance analysis of the gain correction in radiography imaging

Image detectors, which are based on flat panels, suffer from irregular gains due to the non-uniform amplifiers. In order to correct the irregular gains, a gain-correction technique, which is based on the gain map, is conventionally used. By averaging uniformly illuminated images without any objects, the gain map is designed. Increasing the number of images n for averaging provides a good gain-correction performance since the remained quantum noise in the gain map is reduced. In this paper, the performance of the gain-correction is theoretically analyzed based on the signal-to-noise ratio (SNR) with the convergence rate of O (n^-1/2). For real x-ray images, experimental observations are conducted. A filtered gain map, in which a low-pass filter is employed to remove the remained noise, is also observed. SNR as well as the CDRAD contrast-detail phantom are used to experimentally observe the gain-correction performance. It is shown that a 3 × 3 moving average filter can significantly reduce the remained quantum noise.