A specific, learnable software with integrated artificial intelligence (AI) is trained by presenting defined images from a database. The software is taught to pay attention to certain parameters, such as object sizes and perspectives.If the software is confronted with new pictures that do not yet exist in the data base, it examines those for the defined specific values it is trained for. It aligns and calculates how high the likelihood is, that the object is detected correctly. Thus it teaches itself onward, and integrates these new pictures in the data base as well. So the AI is trained to achieve optimal results.
If an arbitrary picture is presented to the AI, it makes a comparison with the pictures of the data base that it is already familiar with.
If the software clearly recognizes what is shown in the picture, or a certain property on which it was trained, this is a match.This can initiate following processes or steer RPA bots.
Here we show you how this self-developed image recognition software with AI can recognize and name a wide variety of objects on images.