Case Study: BSH
This comprises detecting and locating paint flaws when manufacturing fridges.
The flaws might be tiny compared to the size of the surface to be analysed and they might be occasional or appear all over the surface texture. Furthermore, the customer was seeking specific behaviour:
- If the flaw was small enough, it was accepted and should not be considered a flaw.
- In other cases, even though the flaw was very small, it could be considered a defect if it were surrounded by multiple flaws (even if they were equally small).
It was essential to work continuously with the customer to be able to teach the AI to look at the size and the flaws around it.
“When we began working with ENAIA, AI and Deep Learning applied to artificial vision was new technology, yet to demonstrate industrial applications. Now we are convinced that this technology is here to change the industry. We are highly satisfied with our work with ENAIA. There has been fluid two-way feedback, constant reporting on the project which was highly focussed on results and advice that was fundamental on an innovation project such as this. The final result was a very powerful solution, capable of resolving an application that would not have been possible without Deep Learning. This solution will lay the foundations for future applications.”