Image across Domains
This site provides details about my experience at ImageXD v.2.0 (2017), and examples of what we exercised during the event.
Find below the image files needed for Azure-centric activities during tutorial led by Vani Mandava, Director, Data Science Outreach at Microsoft Research at the ImageXD 2017 event. Activities will happen at University of Washington. This exercise will allow practice during ImageXD-2017 core program. All of those who want to use real image data to train machine learning models using Azure tools and will win a $500.00 alocation.
Classify abstract samples into 2 classes
Dataset consists of two classes, containg 2,000 files each, separated in 2 folders (bcc and hcpnone). Each file represents a sample obtained from simulation of a scattering pattern, which produces an image that can be very abstract to non-beamline scientists.
Grazing Incidence Small Angle X-ray Scattering (GISAXS) is a method for characterizing the nanostructural features of materials. This imaging modality supports measurements of the size, shape, and spatial organization of nanoscale objects located on top of thin film materials. Also, GISAXS provide statistical information at the nanometer level. This database is a simulation of GISAXS experiments from a special simulator called HipGISAXS. It is able to handle large input data and perform GISAXS simulations at high resolutions. Using HipGISAXS, we created sample image data with scattering patterns corresponding to two different crystal unit cell structures: BCC: Body Centered Cubic and HCP: Hexagonal Close Packed.
Important references:
A. Hexemer, “HipGISAXS: A massively-parallel highperformance x-ray scattering data analysis code,”
http://www.camera.lbl.gov/gisaxs, 2016, accessed on Dec 13, 2016.
A. Hexemer and P. M¨uller-Buschbaum, “Advanced grazing-incidence techniques for modern soft-matter
materials analysis,” IUCrJ, vol. 2, no. 1, pp. 106–125, 2015.
Having trouble with Pages? Contact Dani and we’ll help you sort it out.