2x2 oversampling in digital radiography imaging for CsI-based scintillator detectors

Dong Sik Kim; Eun Kim; Eunae Lee; Choul Woo Shin

doi:10.1117/12.2253874

Translator Disclaimer

9 March 2017 2x2 oversampling in digital radiography imaging for CsI-based scintillator detectors

Dong Sik Kim, Eun Kim, Eunae Lee, Choul Woo Shin

Author Affiliations +

Proceedings Volume 10132, Medical Imaging 2017: Physics of Medical Imaging; 101323X (2017) https://doi.org/10.1117/12.2253874
Event: SPIE Medical Imaging, 2017, Orlando, Florida, United States

Abstract

In order to efficiently conduct the anti-aliasing filtering in digital radiography imaging, the oversampling scheme using an oversampling detector, in which the sampling frequency is higher than that of the desired detector, is considered in this paper. Instead of using difficult analog anti-aliasing filters, digital anti-aliasing filters are applied to the oversampled data and then their downsampling enables acquiring the desired x-ray images. Supposing an ideal anti-aliasing filtering, the detective quantum efficiency (DQE) performance of the desired detector can be close to that of the oversampling detector since the overlap of the adjacent noise aliases can be minimized while maintaining the frequency amplitude response for the fundamental frequency range. In this paper, a 2 x 2 oversampling is conducted for the desired pixel pitch of 152 μm/pixel and various filters are tested for anti-aliasing filtering. It is shown that securing an enough transition band is important to avoid the ringing artifacts even though the anti-aliasing performance deteriorates due to the wide transition band. From an experiment using a CsI(Tl)-based detector, the aliasing artifact problem is alleviated and a DQE improvement of 0.1 is achieved at 2.5 lp/mm from the oversampling radiography imaging over the binning scheme.

Citation Download Citation

Dong Sik Kim, Eun Kim, Eunae Lee, and Choul Woo Shin "2x2 oversampling in digital radiography imaging for CsI-based scintillator detectors", Proc. SPIE 10132, Medical Imaging 2017: Physics of Medical Imaging, 101323X (9 March 2017); https://doi.org/10.1117/12.2253874

ACCESS THE FULL ARTICLE