skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: A convolutional neural network approach to calibrating the rotation axis for X-ray computed tomography

Abstract

This paper presents an algorithm to calibrate the center-of-rotation for X-ray tomography by using a machine learning approach, the Convolutional Neural Network (CNN). The algorithm shows excellent accuracy from the evaluation of synthetic data with various noise ratios. It is further validated with experimental data of four different shale samples measured at the Advanced Photon Source and at the Swiss Light Source. The results are as good as those determined by visual inspection and show better robustness than conventional methods. CNN has also great potential forreducing or removingother artifacts caused by instrument instability, detector non-linearity,etc. An open-source toolbox, which integrates the CNN methods described in this paper, is freely available through GitHub at tomography/xlearn and can be easily integrated into existing computational pipelines available at various synchrotron facilities. Source code, documentation and information on how to contribute are also provided.

Authors:
; ; ;
Publication Date:
Research Org.:
Argonne National Lab. (ANL), Argonne, IL (United States)
Sponsoring Org.:
USDOE Office of Science (SC), Basic Energy Sciences (BES) (SC-22)
OSTI Identifier:
1354916
DOE Contract Number:
AC02-06CH11357
Resource Type:
Journal Article
Resource Relation:
Journal Name: Journal of Synchrotron Radiation (Online); Journal Volume: 24; Journal Issue: 2
Country of Publication:
United States
Language:
English
Subject:
46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND TECHNOLOGY

Citation Formats

Yang, Xiaogang, De Carlo, Francesco, Phatak, Charudatta, and Gürsoy, Dogˇa. A convolutional neural network approach to calibrating the rotation axis for X-ray computed tomography. United States: N. p., 2017. Web. doi:10.1107/S1600577516020117.
Yang, Xiaogang, De Carlo, Francesco, Phatak, Charudatta, & Gürsoy, Dogˇa. A convolutional neural network approach to calibrating the rotation axis for X-ray computed tomography. United States. doi:10.1107/S1600577516020117.
Yang, Xiaogang, De Carlo, Francesco, Phatak, Charudatta, and Gürsoy, Dogˇa. Tue . "A convolutional neural network approach to calibrating the rotation axis for X-ray computed tomography". United States. doi:10.1107/S1600577516020117.
@article{osti_1354916,
title = {A convolutional neural network approach to calibrating the rotation axis for X-ray computed tomography},
author = {Yang, Xiaogang and De Carlo, Francesco and Phatak, Charudatta and Gürsoy, Dogˇa},
abstractNote = {This paper presents an algorithm to calibrate the center-of-rotation for X-ray tomography by using a machine learning approach, the Convolutional Neural Network (CNN). The algorithm shows excellent accuracy from the evaluation of synthetic data with various noise ratios. It is further validated with experimental data of four different shale samples measured at the Advanced Photon Source and at the Swiss Light Source. The results are as good as those determined by visual inspection and show better robustness than conventional methods. CNN has also great potential forreducing or removingother artifacts caused by instrument instability, detector non-linearity,etc. An open-source toolbox, which integrates the CNN methods described in this paper, is freely available through GitHub at tomography/xlearn and can be easily integrated into existing computational pipelines available at various synchrotron facilities. Source code, documentation and information on how to contribute are also provided.},
doi = {10.1107/S1600577516020117},
journal = {Journal of Synchrotron Radiation (Online)},
number = 2,
volume = 24,
place = {United States},
year = {Tue Jan 24 00:00:00 EST 2017},
month = {Tue Jan 24 00:00:00 EST 2017}
}
  • Synchrotron-based X-ray tomography offers the potential of rapid large-scale reconstructions of the interiors of materials and biological tissue at fine resolution. However, for radiation sensitive samples, there remain fundamental trade-offs between damaging samples during longer acquisition times and reducing signals with shorter acquisition times. We present a deep convolutional neural network (CNN) method that increases the acquired X-ray tomographic signal by at least a factor of 10 during low-dose fast acquisition by improving the quality of recorded projections. Short exposure time projections enhanced with CNN show similar signal to noise ratios as compared with long exposure time projections and muchmore » lower noise and more structural information than low-dose fats acquisition without CNN. We optimized this approach using simulated samples and further validated on experimental nano-computed tomography data of radiation sensitive mouse brains acquired with a transmission X-ray microscopy. We demonstrate that automated algorithms can reliably trace brain structures in datasets collected with low dose-CNN. As a result, this method can be applied to other tomographic or scanning based X-ray imaging techniques and has great potential for studying faster dynamics in specimens.« less
  • A new tool is introduced for screening macromolecular X-ray crystallography diffraction images produced at an X-ray free-electron laser light source. Based on a data-driven deep learning approach, the proposed tool executes a convolutional neural network to detect Bragg spots. Automatic image processing algorithms described can enable the classification of large data sets, acquired under realistic conditions consisting of noisy data with experimental artifacts. Outcomes are compared for different data regimes, including samples from multiple instruments and differing amounts of training data for neural network optimization.
  • There is high demand for efficient, robust and automated routines for tomographic data reduction, particularly for synchrotron data. Registration of the rotation axis in data processing is a critical step affecting the quality of the reconstruction and is not easily implemented with automation. Existing methods for calculating the center of rotation have been reviewed and an improved algorithm to register the rotation axis in tomographic data is presented. The performance of the proposed method is evaluated using synchrotron-based microtomography data on geological samples with and without artificial reduction of the signal-to-noise ratio. The proposed method improves the reconstruction quality bymore » correcting both the tilting error and the translational offset of the rotation axis. The limitation of this promising method is also discussed.« less
  • Pore structure of leached cement pastes (w/c = 0.5) was studied for the first time from micro-scale down to the nano-scale by combining micro- and nano-X-ray computed tomography (micro- and nano-CT). This allowed assessing the 3D heterogeneity of the pore network along the cement profile (from the core to the altered layer) of almost the entire range of cement pore size, i.e. from capillary to gel pores. We successfully quantified an increase of porosity in the altered layer at both resolutions. Porosity is increasing from 1.8 to 6.1% and from 18 to 58% at the micro-(voxel = 1.81 μm) andmore » nano-scale (voxel = 63.5 nm) respectively. The combination of both CT allowed to circumvent weaknesses inherent of both investigation scales. In addition the connectivity and the channel size of the pore network were also evaluated to obtain a complete 3D pore network characterization at both scales.« less
  • Given the importance of the fuel-injection process on the combustion and emissions performance of gasoline direct injected engines, there has been significant recent interest in understanding the fluid dynamics within the injector, particularly around the needle and through the nozzles. Furthermore, the pressure losses and transients that occur in the flow passages above the needle are also of interest. Simulations of these injectors typically use the nominal design geometry, which does not always match the production geometry. Computed tomography (CT) using x-ray and neutron sources can be used to obtain the real geometry from production injectors, but there are trade-offsmore » in using these techniques. X-ray CT provides high resolution, but cannot penetrate through the thicker parts of the injector. Neutron CT has excellent penetrating power but lower resolution. Here, we present results from a joint effort to characterize a gasoline direct injector representative of the Spray G injector as defined by the Engine Combustion Network. High-resolution (1.2 to 3 µm) x-ray CT measurements from the Advanced Photon Source at Argonne National Laboratory were combined with moderate-resolution (40 µm) neutron CT measurements from the High Flux Isotope Reactor at Oak Ridge National Laboratory to generate a complete internal geometry for the injector. This effort combined the strengths of both facilities’ capabilities, with extremely fine spatially resolved features in the nozzles and injector tips and fine resolution of internal features of the needle along the length of injector. Analysis of the resulting surface model of the internal fluid flow volumes of the injector reveals how the internal cross-sectional area and nozzle hole geometry differs slightly from the design dimensions. A simplified numerical simulation of the internal flow shows how deviations from the design geometry can alter the flow inside the sac and holes. Our results of this study will provide computational modelers with very accurate solid and surface models for use in computational fluid dynamics studies and experimentalists with increased insight into the operating characteristics of their injectors.« less