Reading technical diagrams can be tricky due to the number of different symbols, their variations and lack of standardization. Additionally symbols can be placed at any random point, and with context that can change their meaning. 😶‍🌫️

Our #summerinterns applied techniques from computer vision to the problem of reading these technical drawings. They created tools that can find and categorize a variety of symbols from the diagrams, read the adjacent text and match it to the symbols, as well as categorize the symbols further based on the content. 💡

They have created a pipeline that allows the creation of a model for any specific facility with its unique diagrams. This will help us better understand how our facilities are laid out, and eventually create a more standardized way of presenting this information.

The interns have done a great job tackling this difficult problem, and we are excited to see what else they can accomplish in the #future. 🚀

Mathilde Bergenheim, Mikael Medina, Sakarias Frette, Tomine Holgersen Dale and Muhammad Saad Jahangir 🌟

#machinelearning #computervision #summerintern #students #gamechangers