Machine Learning

Data analytics and deep learning applied to health, work and Industry 4.0

"How" and "why" neural networks (so-called "Deep Neural Networks", DNNs) draw certain conclusions and "from what" they generate their knowledge is difficult to understand from the outside. And yet it is precisely this knowledge that forms the basis for important strategic decisions in business processes, in medical diagnostics and therapy (e.g. in the field of pain recognition), but also for the acceptance of such machine learning methods in everyday life (intelligent work tools).

Our researchers are proficient in classical methods of "machine learning" and image analysis as well as in the nowadays often superior methods of "deep learning" and have many years of experience in the development of applications for computer-assisted diagnosis support. This includes the control of microscopes and scanners, automatic analysis up to comprehensible and understandable visualization for the user.

Our "Machine Learning" focus areas

Current & Completed Projects

 

SEMULIN – natural, multimodal interaction for automated driving

Development of a self-supporting natural human-machine interface for automated driving using multimodal input and output modes including facial expressions, gestures, gaze, and speech.

 

TraMeExCo

TraMeExCo (Transparent Medical Expert Companion) is a project funded by Germany’s Federal Ministry of Education and Research (BMBF). Its purpose is to investigate and develop suitable new methods to enable robust and explainable machine learning in complementary applications (digital pathology, pain analysis, cardiology) in the field of medical engineering.

 

Emotion-sensitive robotic platform for therapy

Video-based emotion recognition and multimodal analysis and evaluation of biosignals for better Human-Machine-Interaction

 

Affective Sensing

Cognitive Sensor Technology for Improved Healtcare and Quality of Life

 

Sensor-based AI and satellite-based IoT Communication for Wildlife Research and Conservation

Our goal is to develop camera tags equipped with sensor-based AI and satellite-based IoT communication, intended for bird and wildlife tagging.  

Your advantages - everything from a single source

  • Bundling of core competencies in the fields of medical engineering, image and biosignal analysis, machine learning, software and hardware implementation
  • Many years of experience in building and curating standardized, structured and annotated image and signal databases for training and evaluation of learning procedures (e.g. facial expression and face analysis, biosignals, endoscopy, hematology, pathology, mammography)
  • Long-standing interdisciplinary networking with companies, clinics and universities
  • Sensor (systems) of different physical and technical measurands
  • Modular software kit for the fast realization of a customer-specific data processing chain from acquisition to analysis and visualization of results

Existing infrastructure

  • High Performance Deep Learning 34 node cluster, 72x NVidia P100 GPU, 16x NVidia P40 GPU, 2.8 TB RAM
  • Expandable cinema lab, currently equipped with 70 seats and infrared camera systems
  • Laboratory for microscopy and digital pathology (e.g. Zeiss Axio Imager, Zeiss Axio Scan.Z1, Precipoint M8, immersion microscopy)
  • Endoscopy studio with different rigid and flexible endoscope systems (Wolf, Storz, Schölly, Olympus), light sources, robots and phantoms
  • Electronics laboratory and extensive test equipment 

 

Facial Analysis Solutions

One focus of our research activities is the development of processes for automatically detecting and analyzing objects.  

Digital Health & Analytics

  • Medical Image Analysis
  • Medical Sensors and Analytics
  • Communication and Integrated Care

 

Integrated Sensor Systems

Intelligent sensors are key elements of modern, autonomous systems.