This challenge will have hackers analyze crystallographic nodal data of typical microstructures found in formed metals to attempt to reproduce them with a desired grain size distribution. Hackers will implement RNN, GAN or any AI network to learn the sequential assignment of orientations on the crystallographic nodal graph to reproduce the learned distributions. Typically simulations like these on a large scale can take weeks to months to run, so any gains in performance will be rewarded bonus points!
What it does
Input a phase diagram, output phase analysis. WE WANT TO MAKE A "DATABASE" for all these analyzed composites phase diagrams, in reality, it will make sense for material scientists or anyone to review and add more analysis in future lookups.
How we built it
Backend python, edge recognition, text recognition. Front end Figma mark up --> Html
Challenges we ran into
Phase diagrams need specific attention primarily because of the way the information is embedded into the diagram. The lines in a phase diagram are not of a continuously changing value like in a line plot, but instead represent a boundary. A phase diagram cannot be expressed by a simple table like most line plots, bar charts etc. Further, text could appear in different orientations and associating the text with phase regions (and sometimes vertical lines) is an added complexity that is nontrivial and essential to the final interpretation by material science domain experts. Members worked in different timezones to collaborate on this project :).
Accomplishments that we're proud of
Everything. Backend analysis with labeled regions. Figma wireframes of what the UI would look like.
What's next for GenePhase
Live prediction & prettier frontend. Frontend & backend markup.