problem
Hyundai Elevators sends technicians out to repair elevators. These technicians have to learn to recognize and identify thousands of potential parts from hundreds of elevator models from a catalog that could span decades. This is a very challenging problem that requires dozens of hours of training.
Since it isn’t possible to retroactively barcode every single part in every elevator that has been deployed, we developed an AI model that could learn to recognize parts with images or video taken from a cell phone camera under varied lighting conditions and from many different angles.
By designing a quick and novel data acquisition process we were able to capture all the training data we needed with very little customer time.
use case
Using NeoPulse® AI Studio it was easy to import the data and associated class labels, create an augmented dataset and then train a CNN based image classifier for Hyundai.
Results
NeoPulse® produced impressive results surpassing any expectations. The performance characteristics of the AI model are:
Average precision: 0.994
Average recall: 0.988
Average f1 score: 0.991

We were able to successfully train an AI model to recognize complex industrial parts using Neopulse 3.0 on AWS. The AI solution was built very quickly and was able to recognize objects in unpredictable, real-world environments with high accuracy.


Everybody seems confident in AI, and they actually enjoy solving various AI problems.


The vendor’s services are integral to providing AI solutions for a wider audience. They had an effective project management style, accented by a quick working style.
Dell & AI Dynamics
Learn how we can help enterprises transform global manufacturing operations while addressing the industry’s challenges of an aging workforce and shrinking margins.