34 <strong>E21</strong> <strong>Annual</strong> <strong>Report</strong> <strong>2011</strong>/<strong>2012</strong>
Chapter 4. Radiography and Tomography 35 Scatter Correction Method by Temporal Primary Modulation in X-Ray CT Karsten Schörner 1, 2 , Matthias Goldammer 2 , Jürgen Stephan 2 , Peter Böni 1 1 Physik-Department <strong>E21</strong>, <strong>Technische</strong> Universität München, D-85748 Garching, Germany. 2 Siemens AG, Corporate Technology, D-81739 Munich, Germany. In the last decade, X-ray cone-beam computed tomography (CBCT) has attracted a lot of attention because of its full volume coverage and its high isotropic spatial resolution. Consequentially, the speed-up in scan time has brought CBCT to a wide variety of applications in the medical field as well as in non-destructive testing (industrial CBCT). However, since large volumes are irradiated, a considerable amount of incident radiation is scattered by the patient or object under investigation, but also by environmental structures, e.g. the base of the CT scanner. Compton scattering of X-ray photons is the dominant physical interaction process in typical applications of industrial CBCT. Scattered radiation reaching the detector superimposes on the primary, i.e. non-scattered, radiation and will cause scatter artifacts which severely degrade image quality in CT reconstructions of the volume. Streaks, a general loss of contrast, and an inhomogeneity called cupping artifact belong to the most well-known scatter artifacts. Although different methods and techniques exist, scatter correction is still considered an open problem in CBCT. In the referenced literature [1, 2, 3], we present a novel method for scatter correction in CBCT based on temporal primary modulation. Here, a primary modulator placed between X-ray source and object imprints a spatially varying pattern (e.g. a checkerboard-like pattern) on primary X-rays through attenuation, see Fig. 1. This modulation pattern is only preserved in unaltered primary X-ray photons - scattered X-rays have a broad spatial distribution after the (Compton-) scattering process and, thereby, the original pattern gets lost. Hereby, a distinction between primary and scattered X-rays can be made by appropriate demodulation algorithms. lengths of 1×1mm 2 . The copper fields have a total thickness of 0.7mm which results in 20% beam attenuation. During the CT scan, a motorized linear translation stage moves the primary modulator back and forth by one square length in order to realize the intensity modulation. CT scans of a test phantom are performed: one without scatter correction, and another one employing our new method. The results are shown in Fig. 2: scatter correction by temporal primary modulation leads to significant improvement in image quality (CT slices in middle row) as compared to uncorrected CT slices (top row). Typical scatter artifacts such as streaks (indicated by arrows), loss of contrast as well as the cupping artifact (indicated in line profiles in bottom row) are compensated by the scatter correction method. Figure 2: Axial CT slices and corresponding line profiles. Figure 1: Schematic illustration of temporal primary modulation by shifting the primary modulator back and forth for each CT projection. While primary signals are temporally amplitudemodulated, the scatter fluence is nearly constant. In the first tests of this method, we use a primary modulator with a 2D checkerboard pattern in an industrial CBCT scanner. The primary modulator is fabricated as a ten-layer printed circuit board (PCB) with 99×99 square fields, arranged as 2D checkerboard pattern with bright squares (PCB base material) and dark squares (copper), each with side In summary, the novel scatter correction method based on temporal primary modulation has been successfully tested in first experiments. Our method can easily be applied to existing CBCT scanners and it can be performed during a normal CT-scan, i.e. no additional dose and/or measuring time are necessary for obtaining scatter data. This represents a main advantage over many other scatter correction techniques. While scatter correction by temporal primary modulation was successfully tested here for industrial CBCT, we will explore its application to the medical field in the future. References [1] K. Schörner, M. Goldammer, K. Stierstorfer, J. Stephan, and P. Böni, IEEE Trans. Nucl. Sci. 59, 3278 (<strong>2012</strong>). [2] K. Schörner, M. Goldammer, and J. Stephan, Nucl. Instrum. Meth. Phys. Res. B 269, 292 (<strong>2011</strong>). [3] K. Schörner. PhD thesis, <strong>Technische</strong> Universität München, <strong>2012</strong>.