Detecting Patient Motion in Projection Space for Cone-beam Computed Tomography

Abstract

Abstract. Cone-beam X-Ray systems strictly depend on the imaged object being stationary over the entire acquisition process. Even slight patient motion can affect the quality of the final 3D reconstruction. It would be desirable to be able to discover and model patient motion right from the actual projection images, in order to take it into account during reconstruction. However, while the source-detector arrangement is rotating around the patient, it is difficult to separate this motion from the additional patient motion. We present a novel similarity metric for successive X-Ray projections, which is able to distinguish between the expected rotational and additional sources of motion. We quantitatively evaluate it on simulated dental cone-beam X-Rays and qualitatively demonstrate its performance on real patient data.

Towards geometrical calibration of x-ray computed tomography systems—a review

“…Most imaging methods consist of imaging reference object(s) placed in the instrument's field of view. There are some imaging methods that do not require a dedicated object [38][39][40][41][42][43][44][45][46][47][48][49]; these methods are often known as 'on-line' methods. The guiding principle behind 'on-line' methods is the improvement in image and reconstruction quality rather than determining dimensional accuracy.…”

“…The guiding principle behind 'on-line' methods is the improvement in image and reconstruction quality rather than determining dimensional accuracy. Still, the methods discussed in [38][39][40][41][42][43][44][45][46][47][48][49] can provide insight into geometrical calibration of XCT systems. However, in the interest of conciseness, only imaging methods that utilize dedicated reference objects will be covered in this review.…”

Epipolar Consistency in Transmission Imaging

“…In Computer Vision, it is used, most notably, to estimate the distance of objects to the observer via stereo disparity. While it is in general impossible to estimate the depth of objects from two X-ray images [3], [4], [5], the theoretical model of physics and geometry of transmission imaging still impose certain constraints on X-ray images. This paper illustrates the epipolar geometry of X-ray images and makes the connection to Grangeat's theorem [6], establishing constraints on the information along corresponding epipolar lines.…”

“…However, permission to use this material for any other purposes must be obtained from the IEEE by sending a request to pubs-permissions@ieee.org . purpose [7], [3], but their connection has only recently been established [8]. In contrast to Wein et al [3] or Kyriakou et al [9], the use of Epipolar Consistency Conditions (ECC) [8], [10], [7] does not require intermediate reconstruction and uses a relatively simple and fast metric on pairs of 2D projections.…”

“…purpose [7], [3], but their connection has only recently been established [8]. In contrast to Wein et al [3] or Kyriakou et al [9], the use of Epipolar Consistency Conditions (ECC) [8], [10], [7] does not require intermediate reconstruction and uses a relatively simple and fast metric on pairs of 2D projections. In this paper, we present the derivation of the ECC exploiting synergies in notation of Projective Geometry and tomographic reconstruction.…”