Ease review of the training set from the Quality Center

The elaboration of the training set is a key element of the performance of an inspection application. The training set represents the acceptable variability of your part, and adding a non-compliant part to it will have a negative impact on Spark's sorting performance.

It is therefore important to ensure the quality of the training set make sure there’s no defective parts in it.

This essential check can nevertheless be tedious on Spark:

• Firstly, the production environment is not adapted to this type of task;
• Moreover, the exercise may require the contribution of a quality expert who is not necessarily present on the line.

That's why we now offer the possibility of performing this review from the Quality Center.

From the Qualification page, you will now be able to access your application's training set, review its parts, and request the removal of a part that seems problematic to you.

From Spark, the user will be notified that a suggestion for modifying the application's training set has been made, and the parts to be removed will be visible at the top of the list, labeled as "To Remove". They can then quickly re-evaluate it, decide on its deletion (or not), and train the application at the most opportune moment, considering operational constraints.

<aside> ℹ️

The request to remove from the training set via the Quality Center does not result in any immediate modification of the application on Spark. It is only a suggestion for modification of the training set, and its effective consideration remains in the hands of the user on the line.

</aside>