| About this Abstract |
| Meeting |
MS&T23: Materials Science & Technology
|
| Symposium
|
Materials Informatics for Images and Multi-dimensional Datasets
|
| Presentation Title |
Multi-modal Image Registration for Materials Characterization |
| Author(s) |
Zachary Varley, Marc De Graef, Gregory Rohrer, Megna Shah, Sean Donegan, Michael Uchic |
| On-Site Speaker (Planned) |
Zachary Varley |
| Abstract Scope |
Big data techniques are revolutionizing materials science research by enabling the analysis of large volumes of often disparate data, including images. Image registration is a critical task in many big data pipelines, involving multiple imaging modalities. Traditionally, local optimization approaches have been used for image registration, but we propose a novel approach based on image entropy filters and key point matching. Our key point detector approach is inspired by the difference of Gaussians (DoG) method, while our key point descriptors are based on the scale-invariant feature transform (SIFT). We demonstrate the effectiveness of our approach in the SEM, where planar homographies underpin many registration problems. Our method offers a promising solution for automating the collation of multi-modal data, facilitating informatics-based approaches to improve our understanding of materials structure and properties. |