Quantitative Analysis Methods Using Histogram and Entropy for Detector Performance Evaluation According to the Sensitivi
- PDF / 488,067 Bytes
- 7 Pages / 595.276 x 790.866 pts Page_size
- 103 Downloads / 220 Views
IMAGE & SIGNAL PROCESSING
Quantitative Analysis Methods Using Histogram and Entropy for Detector Performance Evaluation According to the Sensitivity Change of the Automatic Exposure Control in Digital Radiography Jun-Ho Hwang 1,2 & Kyung-Bae Lee 1 & Ji-An Choi 1 & Tae-Soo Lee 2 Received: 27 May 2020 / Accepted: 25 August 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract The purpose of this study is to evaluate detector performance using histogram and entropy analysis according to the sensitivity change of the automatic exposure control (AEC). The experiment was performed as follows: The sensitivity of the detector was analyzed through a normalized histogram with sensitivities of S200, S400, S800, and S1000 of the AEC; the entropy of the image was then analyzed, and the signal volume of the detector was evaluated according to the sensitivity change. As the sensitivity of the AEC was increased from S200 to S1000, the histogram showed underflow, quantization separation, and dynamic range discrepancy. In addition, entropy showed a decrease as sensitivity was set higher; in particular, entropy degradation was more prominent at sensitivities above S800. Through the histogram and entropy analysis, it was concluded that the detector does not reproduce the sensitivity and signal volume accurately when the sensitivity of the AEC is set high in performance evaluation. Keywords Digital radiography (DR) . Automatic exposure control (AEC) . Histogram . Entropy . Detector
Introduction Digital radiography (DR) has been clinically used since the time when the thin film transistor (TFT) technology was applied to the detector. It has been classified as amorphous selenium (a-Se)-based direct digital radiography (DDR) and amorphous silicon (a-Si)-based indirect digital radiography (IDR) [1–3].
* Tae-Soo Lee [email protected] Jun-Ho Hwang [email protected] Kyung-Bae Lee [email protected] Ji-An Choi [email protected] 1
Department of Radiology, Kyung Hee University Medical Center, Seoul 02447, South Korea
2
Department of Biomedical Engineering Graduate School, Chungbuk National University, Cheongju-si 28644, South Korea
Currently, most of the radiation generating devices in hospitals use DR. DR has the advantage that it can acquire images with a wide dynamic range in real time and high resolution by reducing the loss of signals received by the detector, through efficient conversion of visible rays into electrical signals [4–6]. This became the background of various studies that would later provide objective evaluation items to prevent degradation of detector performance [4–6]. A medical image is formed through the acceptance of X-rays that are generated by a radiation generator [4–8], by a suitable detector. It is very important to evaluate detector performance, because it is the detector that eventually converts the X-rays into electrical signals [9, 10]. The main parameters, required to evaluate detector performance and related to image formation are sensitivity, dynamic range, noise,
Data Loading...
