Machine vision is based on extensive image databases that enable the machine to compare it with objects in the outside world (training data) and provide the necessary artificial neural networks. The resulting image models learn basic structures, patterns, colors and objects.

Image models contain the necessary prior knowledge that is used by the computer to identify objects. This prior knowledge is used in a complex training process to adapt existing models forto use new problems. This is also called transfer learning.

The deep learning algorithms used to classify objects compare and classify the individual images to be examined. The deep learning algorithm breaks down the image into a grid and extracts the image information, each examined for a specific image property...


Similar Articles

Connected Bookmarks