The Institute of Signal Processing and System Theory (ISS) at the University of Stuttgart is concerned with methods and applications of signal processing and machine learning (ML). One focus of the methodological development is model-based signal processing such as detection of events, estimation of parameters, fusion of different sensors, identification of systems and blind signal separation. The second focus is data-based machine learning, in particular deep learning. It comprises the design and study of discriminative models (classification, regression), generative models (VAE, GAN) and DNNs for representation learning. Methods for anomaly detection, active learning, transfer learning, continuous learning, semisupervised and unsupervised domain adaptation and causal reasoning are the focus of our investigations.
In addition to the methodological development, the developed procedures are tested and used in various fields of application, e.g. in automotive applications (semantic image segmentation, automotive radar, detection of driver fatigue, ...), medical applications (semantic organ segmentation of CT/MRI images, motion correction with GAN, image-based medical diagnosis, ...), for intelligent systems for the energy revolution (automatic defect diagnosis of solar cells, non-intrusive load monitoring, etc.), for automatic quality prediction in production processes as well for intelligent.test of semiconductor chips and systems.