Improved quality using newly designed algorithms in gamma- and X-ray fusion images with a photon counting CZT detector: Combining the median modified Wiener filter and edge detection method

Abstract

In this study, we designed a photon-counting cadmium zinc telluride (CZT) gamma- and X-ray fusion imaging system and propose a new image processing method to improve image quality. The Geant4 Application for Tomographic Emission (GATE) platform was used to acquire gamma- and X-ray fusion images, and a stack-up parallel-hole collimator, which can greatly improve sensitivity, with a CZT detector was modeled. The newly proposed image processing algorithm to improve the image quality combines the median modified Wiener filter (MMWF), which can efficiently remove noise, and the Canny and Prewitt methods, which are representative edge detection algorithms. As a result, we confirmed that the two new algorithms improve noise and edge information compared to the conventional MMWF. In addition, when comparing the performance of the two proposed algorithms, we demonstrated that when the Canny is fused to MMWF, noise can be removed relatively efficiently, and the Prewitt can obtain improved edge information. The no-reference-based evaluation method was used to test these new algorithms with phantom images, with the same results. Thus, we demonstrate that the method combining MMWF and the edge detection algorithm provides superior image performance to the conventional filtering method in gamma- and X-ray fusion images with photon-counting CZT detectors. © 2021 Elsevier GmbH