Crystal identification in positron emission tomography using nonrigid registration to a Fourier-based template

Abstract

Modern Positron Emission Tomography (PET) detectors typically are made from 2D modular arrays of scintillation crystals. Their characteristic flood field response (or flood histogram) must be segmented in order to correctly determine the crystal of annihilation photon interaction in the system. Crystal identification information thus generated is also needed for accurate system modeling as well as for detailed detector characterization and performance studies. In this paper, we present a semi-automatic general purpose template-guided scheme for segmentation of flood histograms. We first generate a template image that exploits the spatial frequency information in the given flood histogram using Fourier-space analysis. This template image is a lower order approximation of the flood histogram, and can be segmented with horizontal and vertical lines drawn midway between adjacent peaks in the histogram. The template is then registered to the given flood histogram by a diffeomorphic polynomial-based warping scheme that is capable of iteratively minimizing intensity differences. The displacement field thus calculated is applied to the segmentation of the template resulting in a segmentation of the given flood histogram. We evaluate our segmentation scheme for a photomultiplier tube-based PET detector, a detector with readout by a position-sensitive avalanche photodiode (PSAPD) and a detector consisting of a stack of photomultiplier tubes and scintillator arrays. Further, we quantitatively compare the performance of the proposed method to that of a manual segmentation scheme using reconstructed images of a line source phantom. We also present an adaptive method for distortion reduction in flood histograms obtained for PET detectors that use PSAPDs.

An algorithm for automatic crystal identification in pixelated scintillation detectors using thin plate splines and Gaussian mixture models

“…Another flood segmentation method developed by (Chaudhari et al, 2008), involves filtering / identification of the crystals in Fourier space. In this method, intensities across the flood are corrected to be uniform, then a discrete Fourier transform is used to filter the flood histogram and create a template which corresponds to the crystal arrangements of the physical array and is also a lower order spatial approximation of the flood histogram.…”

Spatial distortion correction and crystal identification for position-sensitive avalanche photodiode-based PET scanners

“…d) and (e). The previously proposed adaptive formulae are given in table I, row 3 [16]. The parameter Q in the adaptive formulae is chosen iteratively to maintain a balance between the pincushioning and barrel effects for each device.…”

“…~: (x, y)~(x, y) -u(x, y). The Jacobian of this mapping is given by (c) (b) (a) ( 16) are overlaid on the distortion-corrected flood histograms. Both underlying flood histograms are for the same detector when placed outside and in the magnetic field respectively.…”

“…However, additional compensation is needed in the case of our PETIMRI detectors to account for the asymmetry in the flood histograms caused due to the curvatures of the optical fibers and for the flood histogram rotation in magnetic field. For the distortion-corrected flood histogram, we previously have developed a general purpose semi-automatic segmentation scheme based on Fourier space analysis [16]. In this scheme, we first obtain a template image that exploits the spatial frequency information in the given flood histogram.…”

Positron emission tomography detector performance with different crystal pitches

“…The crystal look-up table has a direct impact on imaging quality [16][17][18]. Currently, there are many kinds of crystal segmentation methods, such as watershed algorithms, Fouriertransform-based segmentation and segmentation algorithms based on data clustering.…”