02/ What can be achieved?
03/ The problem
04/ The opportunity for deep learning
05/ How does the model work?
06/ Data requirements
07/ Test it yourself
08/ Where to learn more?
A model for this use case has been developed using the Peltarion Platform. To help you get started building your own model, we offer a data library with numerous sample datasets, including an image dataset of manufactured parts, the dataset of spectrograms we used for this use case, and an image dataset of cars with different types of damage.
If you’d like to learn how the audio recordings are preprocessed to create spectrograms and then used in a binary image classification model, check out the machinery operating conditions tutorial.
If you're interested in learning more about AI in manufacturing, have a look at the in-depth article about predictive maintenance or read about how one of our customers uses deep learning for manufacturing to improve paper production.