iPattern

Hochschule Niederrhein. Your way.

Institute of Pattern Recognition (iPattern)

Pattern recognition problems frequently arise in the practical application, for example in materials testing, medical diagnosis, the analysis of video and audio data, or in long-term monitoring using event cameras. Methods from the fields of “artificial intelligence” (AI) and “deep learning” are widely used in these contexts. The institute’s name, iPattern, is derived from the English term “pattern recognition”.

We see ourselves as a partner to industry and business, advising you on problems in the field of pattern recognition and developing solutions for them. This can take the form of both direct contract research and work carried out within the framework of publicly funded projects. In addition, we develop and investigate our own research questions in the field of pattern recognition and publish the results as scientific reports and/or as open-source software.

Institute Management

Computer science, graphic data processing

At the Institute for Pattern Recognition, we research and develop methods that enable a computer to perceive information in the same way as a human being. Depending on the type of data being analysed, this gives rise to a variety of applications:

  • Image analysis: recognising objects in images
  • Searching and prediction: Extracting information from large datasets
  • 3D modelling: Generating models by combining data sources
  • Audio signal processing: Recognising speech and speakers
  • Deep learning and artificial intelligence (AI)

For each of these areas of application, we have experts in our team who can advise you on a specific problem and develop a solution.

Image analysis

In this area, we focus on methods for image enhancement, image classification, medical image processing, person tracking in video data, and long-term monitoring – for example, of human activity or biological systems – using event cameras. In addition, our institute develops the open-source software ‘Gamera’, which can be used to create automatic document recognition systems.

Experts:

Computer science, graphic data processing
Mathematics and computer science

Statistical forecasts

In this area, we focus on statistical methods for estimating distributions, identifying correlations and making predictions. These methods are applied in other areas, such as long-term monitoring, or in systems designed to predict future behaviour based on historical data.

Experts:

Mathematics and computer science

3D modelling

In this area, we reconstruct realistic models from geodata. The foundation course includes aerial photographs or point clouds obtained using laser scanners or photogrammetric techniques. The Institute is developing software that automatically generates virtual 3D city models from this data for cadastral and simulation purposes.

Experts:

Mathematics and computer science
Computer science, graphic data processing

Deep Learning and AI

In this area, we focus on the application of methods that use neural networks to make automated decisions and predictions based on large volumes of data. We apply this not only to image and text analysis, but also to the reduction of background noise, as well as to the recognition of content and the speaker from a distorted speech signal.

Expert:

Dr. rer. nat. Michael Gref

Projects funded by the public sector or by foundations

KIRaPol.Next
Artificial intelligence for hazard detection in public spaces using radar systems (01/2026 – 12/2028), funded by the Ministry for Economic Affairs, Industry, Climate Protection and Energy of the State of North Rhine-Westphalia (link to the project homepage)

D3D
Calculation of detailed 3D building models from geodata (05/2026 – 04/2029), funded by the German Research Foundation (DFG) (link to the project homepage)

BeeVision
Monitoring pollinator diversity using dynamic vision sensors (01/2024 – 12/2025), funded by the Carl Zeiss Foundation (link to project homepage)

Sustainable Intensification of Agriculture
Several projects aimed at establishing international collaboration with African researchers (01/2024 – 12/2024), funded by the German Research Foundation (DFG) (link to project homepage)

TUNUKI
Monitoring of tunnel entrances using artificial intelligence (09/2022 – 08/2023), funded by the Federal Ministry of Digital and Transport (mFund Innovation Initiative) (link to project homepage)

Translingo
Digital speech therapy for people who wish to regain their speech following a stroke (people with aphasia) (09/2021 – 08/2024), funded by the Federal Ministry of Education and Research (link to the project homepage)

KiRaPol.5G
Artificial intelligence for radar systems to support police surveillance in public spaces and railway stations (01/2022 – 12/2024), funded by the Ministry of Economic Affairs, Industry, Climate Protection and Energy of the State of North Rhine-Westphalia

MLPCO
Using machine learning to predict crop yields for small-scale farmers (01/2021 – 11/2021), funded by the Federal Ministry of Education and Research

KlaRa
Development of algorithms for object classification using radar signals (09/2020 – 12/2022), funded by the “Central Innovation Programme for SMEs” (ZIM)

plsm
User-centred design of public spaces using sensor technology (04/2018 – 03/2021), funded by the “European Regional Development Fund” (ERDF) (Link to the project homepage)

FahrRad
A sensor-based bicycle assistance system, particularly for older cyclists (05/2017 – 05/2020), funded by the “European Regional Development Fund” (ERDF)


Smart Home Voice Control: Development of a voice-controlled interface for home automation (November 2016 – October 2019), funded by the “Central Innovation Programme for SMEs” (ZIM)

Contract research and development

Voice control of a wet room
Voice control system for a wet room on a train for Siemens Mobility GmbH (2022)

Point cloud analysis
Algorithm for detecting particle trajectories in point clouds from time projection chambers for the University of Santiago de Compostela (2022)

Symmetry scores
Software for calculating symmetry scores of graphical elements for Intelligent Design Automated in Closter, USA (2020)

MRI analysis
Software for semi-automatic muscle segmentation in MRI scans for the ‘German Aerospace Centre’ in Cologne (2019)

Textured 3D city model
Comprehensive 3D city model based on laser scan data and cadastral plans for the city of Dortmund (2019)

Quality control
Automatic detection of solder joint defects for modus hightec-electronics in Willich (2017)

Cost estimation
Method for forecasting repair costs for AUTOonline in Neuss (2014)

 

Independent research projects

Insect monitoring
Automated long-term insect monitoring using dynamic vision sensors to record biodiversity through the use of AI.

3D City Model
Fully automated generation of virtual 3D city models by combining various data sources.
Further information: Online model of Krefeld

Gamera
Python framework for creating automatic document recognition systems.
Further information: Homepage of the open-source software

Wealth distribution
Statistical methods for estimating wealth distribution from incomplete data.
Further information: Technical report & R code

Trajectory Detection
Automatic trajectory detection in point clouds.
Further information: Open-source software & demo

 

Peer-reviewed

Further publications

Regina Pohle-Fröhlich (image analysis)

Christoph Dalitz (Statistical Methods)

Steffen Goebbels (3D models & optimisation)

Michael Gref (Audio Signals & Deep Learning)

Hans-Günter Hirsch (Audio Signal Processing)

Prof. Dr.-Ing. Hans-Günter Hirsch (retired)

Communications, Linear Systems and Networks

hans-guenter.hirsch(at)hs-niederrhein.de

Institute Management

Computer science, graphic data processing
Consulting
Accessibility

HSNR Chatbot