U-SLADS: Unsupervised Learning Approach for Dynamic Dendrite Sampling

TL;DR

This paper introduces an unsupervised learning method using Hierarchical Gaussian Mixture Models for dynamic sampling of dendrite structures in microscopy, enabling faster imaging of metal dendrites during solidification.

Contribution

It presents a novel unsupervised dynamic sampling approach for imaging skeleton-like objects, specifically metal dendrites, which was previously challenging with supervised methods.

Findings

Effective sampling of dendrites in microscopy images.

Reduces data acquisition time and radiation exposure.

Enables fast imaging of primary and secondary dendrite arms.

Abstract

Novel data acquisition schemes have been an emerging need for scanning microscopy based imaging techniques to reduce the time in data acquisition and to minimize probing radiation in sample exposure. Varies sparse sampling schemes have been studied and are ideally suited for such applications where the images can be reconstructed from a sparse set of measurements. Dynamic sparse sampling methods, particularly supervised learning based iterative sampling algorithms, have shown promising results for sampling pixel locations on the edges or boundaries during imaging. However, dynamic sampling for imaging skeleton-like objects such as metal dendrites remains difficult. Here, we address a new unsupervised learning approach using Hierarchical Gaussian Mixture Mod- els (HGMM) to dynamically sample metal dendrites. This technique is very useful if the users are interested in fast imaging the primary and secondary arms of metal dendrites in solidification process in materials science.