Interesting approach back at my first job in Hydrodynamics research BHRA (Now Bhr Group ) in the early 80's

I shared an office with an engineer who's got assigned to research the efficacy of toilets - we had some interesting brainstorming sessions on how to solve some of the experimental problems.

I even started to look at using neural nets to use image recognition to measure how efficient a design was, but as the base hardware cost was over £250,000 that went nowhere, a pity but it would have taken my career in a different direction