Prashant Athavale
Clarkson University
Will present a talk titled:
Denoising and Inpainting of Noisy Grain Orientation Maps
Abstract: Crystallographic orientations can be measured using Scanning Electron Microscope (SEM) based techniques, such as Electron Backscatter Diffraction (EBSD). The orientation data thus obtained may contain noise and missing data.
There are several methods to restore the orientation data. The restorations from these methods may have varying levels of quality. Moreover, many such methods are parameter-dependent. Therefore, finding suitable parameter settings for optimal restorations can take time and effort for users of such methods.
We will discuss an algorithm to obtain high-quality restorations of noisy orientation data and to circumvent the parameter selection problem by automating it. We estimate the noise variance in the data to determine the amount of denoising to apply. We then use this information to determine the stopping criteria for a vector-valued weighted total variation (TV) flow, a nonlinear diffusion applied to the noisy orientation map. We compare the results obtained by our approach with the results from other commonly used denoising filters.
We will also describe our hybrid machine learning-based inpainting algorithm for filling in missing regions in grain orientation data. We start by filling the unknown areas using a partial convolutional network supported by a large set of simulated datasets. Subsequently, these results and known regions are applied as exemplars in Criminisi’s inpainting algorithm to scan the remaining image. We demonstrate that this hybrid technique yields better results than using either approach independently.
Date: October 4, 2024
Location: Snell 212
Time: 11:00am
ZOOM Link for virtual attendance:
https://clarkson.zoom.us/j/93541691606?pwd=cggjBnvRrYbr7mPOCqmmThM0lOOzOd.1
You can access this link by going to the Virtual Class & Recordings tab in Moodle.
Bio: Prashant Athavale earned a Ph.D. in Applied Mathematics and Scientific Computation from the University of Maryland, College Park. He is an Assistant Professor at Clarkson University. He has held prior academic positions at Johns Hopkins University, University of Toronto, and UCLA. His current research interests include mathematical image processing and data science.