Estimation and compensation of the signal-to-noise ratios in radiography imaging under irregular detector gains

Image detectors, which are based on large flat panels, suffer from irregular gains due to the multiple circuits, such as line drives and amplifiers. In this paper, an image formation model, which considers the irregular gains of image detectors, is proposed, and a method for measuring the signal-to-noise ratio (SNR) of the acquired images from the detectors is introduced. The method uses several images taken under the same condition and is theoretically analyzed based on the proposed image formation model. A compensation scheme is then proposed for the method. It is shown that the method is equivalent to measuring the SNR from gain-corrected images, in which gain maps are employed. Despite the irregular gains, we can accurately measure SNRs with the quantum noise by employing the proposed compensation scheme for practical applications of digital radiography detectors.

1. INTRODUCTION

In the plat-panel radiography image detectors, the thin film transistor (TFT) controls the flow of electronic signals of pho-toreceptors, in which the signals are multiplexed, amplified, and digitized by using multiple amplifiers and gate drivers. Since the amplifiers and gate drivers have irregular gains and offsets, acquired images show different gain and offset properties within an image depending on the amplifiers as pixel value differences. Furthermore, variation in photoreceptor size also results in a fixed pattern structure that has a gain component. Hence, the acquired image need to be offset and gain corrected [1], [2]. The offset correction is typically achieved by subtracting a dark image, which is acquired without exposure, from each subsequent exposure image. Conventional gain-correction approach is using an average of multiple uniformly illuminated images as a gain map to prevent the gain-corrected image from being contaminated by the noise of the gain map. The gain correction is important in evaluating the performance of the detector through the signal-to-noise ratio (SNR) and the detective quantum efficiency (DQE) [1], [3].