Knowledge-Based Vision Systems

KVS prioritizes the use of computer vision systems for product and process improvement in industrial applications. Research and development of the KVS group concentrates on:

Biomedical Image Analytics

  • detection and segmenation of cells in microscopic images from histopathologic examinations, but also images of living cells.
  • description and quantification of cell movement (e.g., in the InCeM-project), subcellular structures and other biological objects.

Motion analysis in image sequences for

  • medical-technical purposes (e.g., tracking of cancer cells)
  • analyses in sports (e.g., football, tennis)
  • technical innovations

Quality inspection

  • visual expert systems that involve the emulation of quality analysis by humans
  • especially for metallic, reflective surfaces, industrial textiles and glass
  • measuring and form analysis via registration of 2D and 3D data

KVS pursues a model-based and knowledge-based approach using methods from signal processing, statistics, machine learning (including deep learning (see PDF below)), geometric structure analysis and concepts of formal mathematical logic. Industrial requirements on the one hand and basic research on the other hand inspire the development of novel approaches that lead to patents, scientifically exploitable results, and innovative industrial solutions.

KVS Focus-Profiles

Each topic has a corresponding document for download below, giving a short overview of scope, competences, running projects, and selected publications: