SOLICITUD
interpack_23_reflexx_ai_title_l.jpg

interpack: Enlightening insights into intelligent image evaluation

What is the benefit of artificial intelligence during image evaluation? What advantages does HEUFT reflexx A.I. have in comparison with conventional methods? And what does all this have to do with dancing noodles? There will be elucidation at the beginning of May at the Enlightenment station on our interpack stand.

Have you ever tried to find a tiny piece of spaghetti among a bunch of ring-shaped noodles? Looking for the proverbial needle in a haystack is even easier by comparison! After all, the latter differs from hay in terms of color, whereas the spaghetti differs almost not at all from the other pasta type in terms of coloration and contrast. Thus, particularly when it comes to X-ray shots, classical methods such as gray scale and color recognition cannot be relied upon during the evaluation of detection images of such "structured products".

The spaghetti fragment searched for would be just as difficult to identify as glass splinters, metal particles and other dangerous foreign bodies in the disordered quantity of structured individual products. During image analysis, artificial intelligence is thus required to pay precise attention to the shape in order to reliably distinguish straight noodles from curly ones – as well as O-rings, nuts and other very similarly shaped objects. We will be demonstrating exactly how this works and what smart technologies we have developed for this purpose at one of the Enlightenment stations at our interpack trade show booth:

AI for precise shape recognition

We make curly noodles dance on a vibrating plate; without orientation, they move back and forth there in a crisscross pattern. In this way you can try out for yourself what happens when a particularly shaped spaghetti or even a very similarly shaped O-ring is added: HEUFT reflexx A.I. finds even the smallest deviations from the original shape and structure of the good products on the live images – using artificial intelligence!

We will be pleased to demonstrate to you at our Enlightenment station at the interpack why the use of multi-layered neural networks by means of Deep Learning is the suitable procedure here, unlike for example during the neck finish inspection of glass containers, how exactly this works and what benefits this brings in particular during the evaluation of X-ray images. At booth A43 in hall 13, you may not find the needle in the haystack, but you will definitely find one straight among numerous curled noodles!