Image-Based Experimental Analysis 

D. Ushizima, H. Krishnan and T. Perciano, M. Haranczyk

Designing, Building and Optimizing Image Analysis of Diverse Materials

Algorithms are not keeping up with the rapid increase in the capabilities of imaging sensors. Many DOE research laboratories store digital images as part of their experimental records. Limitations in image analysis hamper our ability to understand the data acquired by high resolution sensors. As an example, much of the data acquired at imaging facilities is manually inspected, delaying access to experimental results. Invaluable information, encoded in these large datasets and obtained at considerable cost, is often lost. Currently, users are forced to utilize memory-bound tools that require drastic downsampling in order to analyze overwhelming data sizes/rates. Much of the precision and nuance captured by experimental apparatus vanishes with improper downsampling. Analysis of data coming from high resolution, high-throughput sensors is a fundamental challenge for data-intensive science. Our focus has been on building techniques for compressing large data, devising comparison metrics, and inventing algorithms to guide and optimize experiments. Advances in image-based methods will save time between experiments, make efficient use of materials, and open up imaging instruments to more experiments for more users. Some of the applications using image-based experiments are:

STEM Microscopy

Films for future integrated circuits

 

Porous materials are responsible for insulating the interconnects between transistors. The architecture of periodic mesoporous organosilicas (PMO) can be controlled by chemical agents, which can tune pore dimensions, wall thicknesses and overall porosity. Quantification of film structures organization is key to a successful film, and our team is carrying out a more systematic study than has been done before.

A key challenge is to measure order from scanning transmission electron microscopy (STEM).  

 

Paper published in Advanced Function Materials (DOI: 10.1002/adfm.201501059)

Collaboration: NCEM, Molecular Foundry

High-resistance ceramic composites under strain

 

Aiming at detecting and quantifying composite cracks and fiber breaks, we developed filtering tools to implement recognition algorithms that handles large  image  stacks. This new  image  processing toolkit works as a plugin for free software ImageJ/Fiji. F3D  includes non-linear filters, such as mathematical morphology (MM) operators, bi- lateral and  median  filters,  implemented in OpenCL to take  advantage of GPUs.

 

Collaboration: MSD, ALS

 

X-Ray Micro-CT

Geological Samples

 

Aiming at detecting and quantifying composite cracks and fiber breaks, we developed filtering tools to implement recognition algorithms that handles large  image  stacks. This new  image  processing toolkit works as a plugin for free software ImageJ/Fiji. F3D  includes non-linear filters, such as mathematical morphology (MM) operators, bi- lateral and  median  filters,  implemented in OpenCL to take  advantage of GPUs.

 

Collaboration: MSD, ALS

 

Scanning Electron Microscopy (SEM)

 

Toxicity of Nanoparticles

 

Our goal is to quantify the risks posed by engineered nanoparticles, particularly hydroxyapatites. This project investigates computational tools based on computer vision algorithms applied to SEM images, to predict toxicity of nanoparticles based on their morphology among other structural/physical properties. We have analyzed SEM images to obtain particle information (particle shape/size/surface/etc), and are building a statistical model that correlates both nano and microscale parameters with a biological response to those particles (e.g. toxicity).

 

Collaboration: European Commission NanoBridges Program (Marie Curie IRSES)

Simulations

 

Filamentous Networks

 

Filaments are fundamental structures permeating materials that define a channel network which influences material porosity and crack development. Markov random fields provide a consistent way to model context dependent entities as image voxels, by characterizing  mutual influences between these entities. Our previous work uses these ideas to detect thin structures from 2D images. We are now working on new algorithms adapted to 3D structures.

 

Collaboration: within CRD