"U.EDU: Unified Education - Media education along the teacher training chain" (BMBF funding code: 01JA2029)
In the TransferLAB project, M.Sc. Johannes Bayer is further developing the CircuitGraph project software based on existing research results. This combines mechanisms for the training and inference of models for image recognition and the explanation of individual symbols using semantic technologies with a graphical user interface for displaying the images and overlaying the extracted graph structures and explanations. A major challenge was the creation of training material, which, like the software itself, was made publicly available.
Graph Extraction and Knowledge Inference for Circuit Diagram Images
CircuitGraph is a software for extracting electrical graph structures from handwritten and printed circuit diagram images using machine learning-based image processing as well as identifying high-level engineering concepts in those circuits and helping the user by automatically completing those structures.
CircuitGraph is intended to be used by electrician students to help them in understanding circuits as well as their teachers in assessing their performances.
