SLShane Jamie Latham

Shane Jamie Latham

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Shane Jamie Latham has a total of 45 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Techniques for high-fidelity X-ray micro-tomography of additively manufactured metal components

    by A.M. Kingston, Q. Yang, M.G. Grewar, O. Delgado-Friedrichs, G.R. Myers, S.J. Latham, A.P. Sheppard

    Additive manufacturing is revolutionising the production of complex-shaped components and finding widespread application in the industries where production volumes are low and part geometries are complex. High-fidelity 3D imaging of additively manufactured metal parts, as the most effective nondestructive evaluation method, is crucial to maintaining the quality of safety-critical components during manufacturing. However, 3D X-ray imaging of large metallic parts is complicated by strong, nonlinear interactions between the sample and the X-ray beam; phenomena such as beam hardening, photon starvation and X-ray scatter are difficult to account for and cause artefacts in the computed 3D tomographic volume. These artefacts can obscure real defects and cause severe inaccuracies in analysis and evaluation. Here we present several techniques developed in the CTLab at the Australian National University that enhance the micro-tomography instruments’ ability to image large, complex metallic components.


  • Open Access

    PI-Line Difference for Alignment and Motion-Correction of Cone-Beam Helical-Trajectory Micro-Tomography Data

    by Olaf Delgado-Friedrichs, Andrew Maurice Kingston, Shane J. Latham, Glenn R. Myers, Adrian P. Sheppard

    In micro-tomography, variants of helical X-ray source trajectories, e.g., double-helix (DH) and space-filling (SF), are attractive alternatives to the conventional circular trajectory, as they satisfy data-sufficiency conditions; this enables exact reconstruction, and large cone-angle (or high flux) imaging. Geometric alignment of micro-tomography experimental data, i.e., radiographs or projection data, to the required micron precision is a difficult problem. Here we consider criteria based on differences in attenuation along opposing rays as a postacquisition software alignment technique. These opposing rays are called PI-lines and lie on lines that intersect two points of the scanning trajectory and pass through the region of support. The PI-line difference method is particularly appealing due to its low computational cost and small set of inherent assumptions, however, previous studies have exposed some limitations in precision. The number and distribution of PI-lines is highly dependent on the trajectory and thus so is the robustness of PI-line difference; here we show that DH and SF trajectories are particularly amenable to this technique. For these trajectories, we observe that the technique is applicable to both static and per-projection alignment estimation. We present results where PI-line difference alignment estimates are of equivalent accuracy as alignment estimates obtained from existing state-of-the-art methods: a tomogram sharpness method for static alignment and a re-projection-alignment method for per-projection alignment. In both cases, the computational expense of PI-line difference alignment estimation is a fraction of the tomogram-based methods.


  • Open Access

    Space-Filling X-Ray Source Trajectories for Efficient Scanning in Large-Angle Cone-Beam Computed Tomography

    by Andrew Maurice Kingston, Glenn R. Myers, Shane J. Latham, Benoit Recur, Heyang Li, Adrian P. Sheppard

    We present a new family of X-ray source scanning trajectories for large-angle cone-beam computed tomography. Traditional scanning trajectories are described by continuous paths through space, e.g., circles, saddles, or helices, with a large degree of redundant information in adjacent projection images. Here, we consider discrete trajectories as a set of points that uniformly sample the entire space of possible source positions, i.e., a "space-filling" trajectory (SFT). We numerically demonstrate the advantageous properties of the SFT when compared with circular and helical trajectories: a) the most isotropic sampling of the data, b) optimal level of mutually independent data, and c) an improved condition number of the tomographic inverse problem. The practical implications of these properties in tomography are also illustrated by simulation. We show the SFT provides greater data acquisition efficiency, and reduced reconstruction artifacts when compared with helical trajectory. It also possesses an effective pre-conditioner for fast iterative tomographic reconstruction.


  • Reprojection Alignment for Trajectory Perturbation Estimation in Microtomography

    by Shane J. Latham, Andrew Maurice Kingston, Benoit Recur, Glenn R. Myers, Olaf Delgado-Friedrichs, Adrian P. Sheppard

    For standard laboratory micro-tomography systems, acquired radiographs do not always adhere to the strict geometrical assumptions of the reconstruction algorithm. The consequence of this geometrical inconsistency is that the reconstructed tomogram contains motion artefacts, e.g. blurring, streaking, double-edges. To achieve a motion-artefact-free tomographic reconstruction, one must estimate, and subsequently correct for, the per-radiograph experimental geometry parameters. In this paper we examine the use of re-projection alignment (RA) to estimate per-radiograph geometry. Our simulations evaluate how the convergence properties of RA vary with: motion-type (smooth versus random), trajectory (helical versus discrete-sampling 'space-filling' trajectories) and tomogram resolution. The idealized simulations demonstrate for the space-filling trajectory that RA convergence rate and accuracy is invariant with regard to the motion-type and that the per-projection motions can be estimated to less than 0.25 pixel mean absolute error by performing a single quarter-resolution RA iteration followed by a single half-resolution RA iteration. The direct impact is that, for the space-filling trajectory, one can incorporate RA in an iterative multi-grid reconstruction scheme with only a single RA iteration per multi-grid resolution step. We also find that for either trajectory, slowly varying vertical errors cannot be reliably estimated by employing the RA method alone; such errors are indistinguishable from a trajectory of different pitch. This has minimal effect in practice because RA can be combined with reference frame correction which is effective for correcting low-frequency errors.


  • Open Access

    array_split: Multi-dimensional array partitioning

    by Shane J Latham

    The array_split (Latham 2017) Python package extends existing dense array partitioning capabilities found in the numpy (Walt, Colbert, and Varoquaux 2011) (numpy.array_split) and skimage (Van der Walt et al. 2014) (skimage.util.view_as_blocks) Python packages. In particular, it provides the means for partitioning based on array shape (rather than requiring an actual numpy.ndarray object) and can partition into sub-arrays based on a variety of criteria including: per-axis number of partitions, total number of sub-arrays (with per-axis number of partition constraints), explicit sub-array shape and constraining a partitioning with an upper bound on the resulting sub-array number of bytes. Application areas include: Parallel Processing Data parallelism by partitioning array for multi-process concurrency (e.g. multiprocessing ("Multiprocessing-Process-Based Parallelism" 2017) or mpi4py (Dalcin et al. 2011)) based on number of cores, or partitioning for accelerator hardware concurrency (e.g. pyopencl or pycuda [kloeckner_pycuda_2012]) based on hardware memory limits. File I/O Partitioning large arrays for output to separate files (e.g. as part of a virtual dataset (The HDF Group 1997-1997-NNNN, Collette (2013))) based on maximum file size, or out-of-core partitioning based on in-core memory limits.

  • The index of dispersion as a metric of quanta – unravelling the Fano factor

    by Wilfred K. Fullagar, Mahsa Paziresh, Shane J. Latham, Glenn R. Myers, Andrew M. Kingston

    In statistics, the index of dispersion (or variance-to-mean ratio) is unity (σ2/〈x〉 = 1) for a Poisson-distributed process with variance σ2for a variablexthat manifests as unit increments. Wherexis a measure of some phenomenon, the index takes on a value proportional to the quanta that constitute the phenomenon. That outcome might thus be anticipated to apply for an enormously wide variety of applied measurements of quantum phenomena. However, in a photon-energy proportional radiation detector, a set ofMwitnessed Poisson-distributed measurements {W1,W2,…WM} scaled so that the ideal expectation value of the quantum is unity, is generally observed to give σ2/〈W〉 < 1 because of detector losses as broadly indicated by Fano [Phys. Rev.(1947),72, 26]. In other cases where there is spectral dispersion, σ2/〈W〉 > 1. Here these situations are examined analytically, in Monte Carlo simulations, and experimentally. The efforts reveal a powerful metric of quanta broadly associated with such measurements, where the extension has been made to polychromatic and lossy situations. In doing so, the index of dispersion's variously established yet curiously overlooked role as a metric of underlying quanta is indicated. The work's X-ray aspects have very diverse utility and have begun to find applications in radiography and tomography, where the ability to extract spectral information from conventional intensity detectors enables a superior level of material and source characterization.


  • Tomography of atomic number and density of materials using dual-energy imaging and the Alvarez and Macovski attenuation model

    by M. Paziresh, A. M. Kingston, S. J. Latham, W. K. Fullagar, G. M. Myers

    Dual-energy computed tomography and the Alvarez and Macovski [Phys. Med. Biol. 21, 733 (1976)] transmitted intensity (AMTI) model were used in this study to estimate the maps of density (ρ) and atomic number (Z) of mineralogical samples. In this method, the attenuation coefficients are represented [Alvarez and Macovski, Phys. Med. Biol. 21, 733 (1976)] in the form of the two most important interactions of X-rays with atoms that is, photoelectric absorption (PE) and Compton scattering (CS). This enables material discrimination as PE and CS are, respectively, dependent on the atomic number (Z) and density (ρ) of materials [Alvarez and Macovski, Phys. Med. Biol. 21, 733 (1976)]. Dual-energy imaging is able to identify sample materials even if the materials have similar attenuation coefficients at single-energy spectrum. We use the full model rather than applying one of several applied simplified forms [Alvarez and Macovski, Phys. Med. Biol. 21, 733 (1976); Siddiqui et al., SPE Annual Technical Conference and Exhibition (Society of Petroleum Engineers, 2004); Derzhi, U.S. patent application 13/527,660 (2012); Heismann et al., J. Appl. Phys. 94, 2073–2079 (2003); Park and Kim, J. Korean Phys. Soc. 59, 2709 (2011); Abudurexiti et al., Radiol. Phys. Technol. 3, 127–135 (2010); and Kaewkhao et al., J. Quant. Spectrosc. Radiat. Transfer 109, 1260–1265 (2008)]. This paper describes the tomographic reconstruction of ρ and Z maps of mineralogical samples using the AMTI model. The full model requires precise knowledge of the X-ray energy spectra and calibration of PE and CS constants and exponents of atomic number and energy that were estimated based on fits to simulations and calibration measurements. The estimated ρ and Z images of the samples used in this paper yield average relative errors of 2.62% and 1.19% and maximum relative errors of 2.64% and 7.85%, respectively. Furthermore, we demonstrate that the method accounts for the beam hardening effect in density (ρ) and atomic number (Z) reconstructions to a significant extent.