Robert Boschgmbh

Fraunhofer-gesellschaft Zur Foerderung Der Angewandten Forschung E.v.

Karlsruher Institut Fur Technologie

Eduardo Monari

Matthias Hartrumpf

Robin Gruna

Thomas Längle

Henning Schulte

Sabrina Hoppe

Michael Heizmann

Ruyu Wang

Markus Vogelbacher

A computer-implemented method of training a generator for transforming a given image according to a given target foreground domain is disclosed. A generator, discriminator, Mapping network, and a Style-Content encoder are trained. Furthermore, a method of image to image translation according to a given target foreground domain by the trained generator is provided. The generator may be trained to generate images depicting defects.

Device and method to improve synthetic image generation of image-to-image translation of industrial images

The invention relates to an apparatus for controlling the quality of transparent objects, comprising at least one light source being adapted to illuminate the object, and at least one image capturing device, by means of which at least one image of the object can be detected, wherein the apparatus comprises further at least one reflector, wherein the reflector and the image capturing device define a beam path, wherein the reflector, the image capturing device, and the light source are arranged such that the object can be arranged between the reflector, on the one hand, and the image capturing device and the light source, on the other hand, such that a single sheet of material is located in the beam path. Furthermore, the invention relates to a corresponding method for controlling the quality of transparent objects.

Apparatus and method for controlling the quality of transparent objects

Growing community of inventors

Karlsruhe, Germany

Eduardo Monari