Unsupervised Deep Learning for Laboratory-Based Diffraction Contrast Tomography
- PDF / 1,282,816 Bytes
- 7 Pages / 595.276 x 790.866 pts Page_size
- 57 Downloads / 168 Views
THEMATIC SECTION: 5TH INTERNATIONAL CONGRESS ON 3D MATERIALS SCIENCE
Unsupervised Deep Learning for Laboratory‑Based Diffraction Contrast Tomography Emil Hovad1 · Haixing Fang2 · Yubin Zhang2 · Line Katrine Harder Clemmensen1 · Bjarne Kjær Ersbøll1 · Dorte Juul Jensen2 Received: 19 June 2020 / Accepted: 3 October 2020 © The Minerals, Metals & Materials Society 2020
Abstract An important leap forward for the 3D community is the possibility to perform non-destructive 3D microstructural imaging in the home laboratories. This possibility is profiled by a recently developed technique—laboratory X-ray diffraction contrast tomography (LabDCT). As diffraction spots in LabDCT images are the basis for 3D reconstruction of microstructures, it is critical to get their identification as precise as possible. In the present work we use a deep learning (DL) routine to optimize the identification of the spots. It is shown that by adding an artificial simple constant background noise to a series of forward simulated LabDCT diffraction images, DL can be trained and then learn to remove high frequency noise and low frequency radial gradients in brightness in the real experimental LabDCT images. The training of the DL routine is unsupervised in the sense that no human intervention is needed for labelling the data. The reduction in high frequency noise and low frequency radial gradients in brightness is demonstrated by comparing line profile scans through the experimental and the DL output images. Finally, the implications of this reduction procedure on the spot identification are analysed and possible improvements are discussed. Keywords LabDCT · Deep learning · Noise reduction · Diffraction images · 3D microstructures
Introduction
* Emil Hovad [email protected] Haixing Fang [email protected] Yubin Zhang [email protected] Line Katrine Harder Clemmensen [email protected] Bjarne Kjær Ersbøll [email protected] Dorte Juul Jensen [email protected] 1
DTU Compute, Department of Applied Mathematics and Computer Science, Technical University of Denmark, Richard Petersens Plads, Building 324, 2800 Kgs, Lyngby, Denmark
DTU Mechanical Engineering, Department of Mechanical Engineering, Technical University of Denmark, Produktionstorvet, Building 425,2800 Kgs, Lyngby, Denmark
2
Non-destructive 3D characterization of materials microstructures using methods at large international synchrotron facilities offers substantial advantages compared to conventional 2D methods. For metals and alloys in particular, the possibility to follow the microstructural evolution in the bulk has led to several major scientific breakthroughs [1–5]. However, to broaden the use of this novel and outstanding characterization possibility, there is a need for techniques that can operate in the home laboratories. Laboratory X-ray diffraction contrast tomography (LabDCT) offers such opportunities [6, 7]. A cornerstone in the LabDCT method is precise identification of the diffraction spots. In the commercially available LabDCT software, the images of LabDCT projections are normally
Data Loading...
