ProgramMaster Logo
Conference Tools for 2016 TMS Annual Meeting & Exhibition
Login
Register as a New User
Help
Submit An Abstract
Propose A Symposium
Presenter/Author Tools
Organizer/Editor Tools
About this Abstract
Meeting 2016 TMS Annual Meeting & Exhibition
Symposium Driving Discovery: Integration of Multi-Modal Imaging and Data Analysis
Presentation Title Bingham Mixture Model for Efficient Microtexture Estimation from Discrete Orientation Data
Author(s) Stephen Niezgoda, Eric Magnuson
On-Site Speaker (Planned) Stephen Niezgoda
Abstract Scope The proper analysis of complex and multi-modal imaging requires careful consideration of the nature and properties of the collected data. Often techniques that provide no information-theoretic guarantees as to their optimality in describing the given data set and are prone to overfitting are adopted. The analysis of spatially resolved discrete orientation measurements, as produced by EBSD, serves as a ubiquitous example. The estimation of orientation distributions via generalized spherical harmonic expansion relies on ad hoc methods for choosing parameters and enforcing constraints. Here we present an unsupervised learning approach for the estimation of orientation distributions as a finite mixture of Bingham distributions. The Bingham distribution is the maximum entropy distribution for the rotation group. The algorithm introduces a minimum message length criterion, to balance data likelihood against model complexity. This approach leads to ODFs which are less likely to overfit the data, eliminating the need for a priori parameter choices.
Proceedings Inclusion? Planned: A print-only volume

OTHER PAPERS PLANNED FOR THIS SYMPOSIUM

3D and 4D Characterization of Failure Mechanisms in Commercial Li-Ion Batteries
Bingham Mixture Model for Efficient Microtexture Estimation from Discrete Orientation Data
Correlation of Multi-modal Chemical Imaging with Computational Simulations for Energy Materials
Digital Representation of Materials Grain Structure from Four-Dimensional X-ray Microtomography Data
Error Analysis of Near-field High Energy Diffraction Microscopy
In Situ Synchrotron Quantification of Evolving Solidification Microstructures in Ni and Co Based Alloys
Integrated Imaging: The Sum is Greater than the Parts
Integrated Multimodal Imaging of Cathodes for Lithium Ion Battery
Methodology for Reconstruction of Samples Analyzed with Simultaneous Neutron and X-Ray Imaging
Modeling Multi-modal Images of Photocatalysis on Cu2O
Multi-Modality Imaging at the Hard X-ray Nanoprobe Beamline at the NSLS-II
Multi-scale, Multi-Model Analysis of Deformation Behavior in Metallic Materials by X-ray Microtomography, FIB, and EBSD
Neutrons, Materials and Data Challenges
Real Time Analysis, Interpretation and Experimental Steering for Electron Microscopy
Recognizing Patterns from Experimental Data
Structure Quantification, Property Prediction and 4D Reconstruction Using Limited X-ray Tomography Data

Questions about ProgramMaster? Contact programming@programmaster.org