Logo der Physikalisch-Technischen Bundesanstalt

WORKSHOP Machine learning in optical analytics

Machine learning in optical analytics encompasses a wide range of applications in which optical instruments and analysis methods are improved by self-learning algorithms. Application examples include the quantitative determination of material composition from complex spectra, fast real-time data analysis incorporating prior knowledge, or intelligent alarm triggers. Current challenges lie in the extension of the application scenarios for machine learning in optical analysis and the creation of a data basis for training the algorithms. Although machine learning is highly developed in some areas of optics such as image processing, there is great potential for novel applications using the unique capabilities and big data offered by optical and X-ray analytics.

For users of optical analytics this workshop will show how machine learning enables completely new analysis methods, data fusion from several methods, as well as fast processing of big data not available with traditional algorithms. For practitioners in machine learning, the conference may trigger interest in the use cases and rich data of advanced optical analytics, so that inspiring joint developments will be triggered.

Machine Learning for Spectroscopic Data

PD Dr. Thomas Bocklitz (Friedrich-Schiller-University and IPHT)

Real-time Dynamic Strain Sensing in Optical Fibers Using Artificial Neural Networks

Dr. Sascha Liehr (BAM)

Artificial Intelligence in Surface Analysis

Dr. habil. Thorsten Kampen (SPECS GmbH)

FT-IR Hyperspectral Imaging and Artificial Neural Network Analysis for Rapid Identification of Pathogenic Bacteria

Dr. Peter Lasch, (Robert-Koch-Institut)

Interpretable & Transparent Deep Learning
Dr. Wojciech Samek (Fraunhofer HHI)

Challenges in Machine Learning and Computer Vision at Amazon

Dr. Tammo Rukat (Amazon)

Machine Vision in Production and Logistics
Johannes Hügle (Fraunhofer IPK)