Thomas Zielke

Hochschule Düsseldorf · Faculty of Mechanical and Process Engineering

Soft Sensors are predictors for measurements that are difficult or impossible to obtain by means of a physical sensor. A soft sensor delivers virtual measurements based on several or many physical measurements and a mathematical or numerical model that incorporates the physical knowledge about the interdependencies. In industrial processes, control...

The integration of AI components into products and industrial processes is currently limited to industries that do not have requirements for rigorous software verification, as specified by ISO 26262 for the functional safety of road vehicles. AI systems based on Machine Learning (ML) typically deal with uncertain data in a context that bears uncert...

Dieses Dokument gibt die Anforderungen an, unter denen Bilderkennungsprobleme in der Medizin mit Hilfe eines Deep-Learning-Bilderkennungssystems bearbeitet werden können. Es erlaubt Entscheidungsträgern, Kenntnisse über die Anwendungsmöglichkeiten eines Deep-Learning-Bilderkennungssystems in der Medizin und seine Struktur zu gewinnen. Mit Hilfe die...

This paper has three intertwined goals. The first is to introduce a new similarity measure for scatter plots. It uses Delaunay triangulations to compare two scatter plots regarding their relative positioning of clusters. The second is to apply this measure for the robustness assessment of a recent deep neural network (DNN) approach to dimensionalit...

Many standards development organizations worldwide work on norms for Artificial Intelligence (AI) technologies and AI related processes. At the same time, many governments and companies massively invest in research on AI. It may be asked if AI research has already produced mature technologies and if this field is ready for standardization. This art...

Many standards development organizations worldwide work on norms for Artificial Intelligence (AI) technologies and AI related processes. At the same time, many governments and companies massively invest in research on AI. It may be asked if AI research has already produced mature technologies and if this field is ready for standardization. This art...

With the completion of their bachelor thesis, engineering students have usually carried out and compiled a genuine piece of research. However, even with a thorough supervision, the exposition of results and arguments often remains limited to a small internal circle of stakeholders in a protected research environment and/or in an examination process...

Deep Learning (DL) bildet den neusten State-of-the-Art der Bilderkennung und wird mit Erfolg in vielen Gebieten angewendet, z.B. für autonomes Fahren, Gesichtserkennung und Spracherkennung[1]. Deep Learning ist eine Methode der künstlichen Intelligenz (KI) und basiert auf tiefen, künstlichen neuronalen Netzen. Anstatt mit viel Aufwand fest codierte...

We elaborate on the robustness assessment of a deep neural network (DNN) approach to dimensionality reduction for data visualization. The proposed DNN seeks to improve the class separability and compactness in a low-dimensional feature space, which is a natural strategy to obtain well-clustered visualizations. It consists of a DNN-based nonlinear g...

Die quantitative Analyse von Studierendendaten aus einem Campusmanagementsystem ist ein Instrument der Qualitätssicherung und der administrativen Planung. Daten aus den individuellen Studienverläufen lassen sich zukünftig auch dafür verwenden, Maßnahmen zur Verbesserung des Studienerfolgs und zur Verringerung der Abbruchquote zu entwickeln und zu e...

Deriving mathematical models for the simulation of dynamic processes is costly and time-consuming. This paper examines the possibilities of deep neural networks (DNNs) as a means to facilitate and accelerate this step in development. DNNs are machine learning models that have become a state-of-the-art solution to a wide range of data analysis and p...

We present an approach to visualizing gradient descent methods and discuss its application in the context of deep neural network (DNN) training. The result is a novel type of training error curve (a) that allows for an exploration of each individual gradient descent iteration at line search level; (b) that reflects how a DNN’s training error varies...

Undergraduate students in engineering often have little exposure to the world of scientific publishing and the culture of sharing research work. While the beneficial exchange between research and teaching is well promoted, while the benefits of conference participation for students have already been surveyed and while courses on scientific writing...

This article looks at useful applications of Grading Analytics, i.e. the analysis of data in a conventional student information system of a university. Much research in educational data mining focuses on Learning Analytics, a field that heavily exploits data that comes from digital footprints of students as they use online learning tools and learni...

– This article is on a university course called Engineering Conferences. It has been developed by the authors and installed as a mandatory part of the curriculum in three master’s degree programs for engineering students. The participants of the course are postgraduates with different nationalities, mostly German, and different technical background...

Die Entwicklung mathematischer Computermodelle zur Simulation dynamischer Prozesse ist aufwendig und zeitintensiv. Um diesen Vorgang zu beschleunigen und zu vereinfachen werden im vorliegenden Beitrag die Möglichkeiten selbstlernender Verfahren aufgezeigt, vergleichbare Modelle zu erstellen. Eingesetzt werden Deep Neural Networks (DNNs), eine spezi...

erschienen in: Die Neue Hochschule, Heft 3/2017

This article is on a course called Engineering Conferences which the authors have developed and installed as a mandatory part of the curriculum in three master's degree programs for engineering students. The participants of the course are postgraduates with different nationalities, mostly German, and different technical backgrounds. They study Mech...

A course called "Engineering Conferences" is presented that the authors have developed and installed as a mandatory part of the curriculum in Master programs for engineering students. The idea is to go beyond teaching the standards of academic writing and skills for working with scientific publications. By using a learner-centered approach, we get...

This poster summarizes a few facts, figures and resources related to the Master course ”Engineering Conferences”.

The event ”Light on” brings together researchers from all the different faculties of the Düsseldorf University of Applied Sciences. It provides a platform for presenting current research work and ideas for future work. This short talk sketches the idea of transforming reflected light, as captured by a camera, into decisions or into actions. While t...

We present an approach to feature extraction that is a generalization of the classical linear discriminant analysis (LDA) on the basis of deep neural networks (DNNs). As for LDA, discriminative features generated from independent Gaussian class conditionals are assumed. This modeling has the advantages that the intrinsic dimensionality of the featu...

Deep Neural Networks (DNNs) denote multilayer artificial neural networks with more than one hidden layer and millions of free parameters. We propose a Generalized Discriminant Analysis (GerDA) based on DNNs to learn discriminative features of low dimension optimized with respect to a fast classification from a large set of acoustic features for emo...

Face detection on time-of-flight image data is a challenging pattern recognition task because it involves the combination of information from two different sensor modalities. In a 2009 paper Boehme et al. show how this task can be solved by a cascade of boosted classifiers that are constructed by a modified Viola-Jones algorithm. Contrary to this a...

We propose a framework for optimizing Deep Neural Networks (DNN) with the objective of learning low-dimensional discriminative features from high-dimensional complex patterns. In a two-stage process that effectively implements a Nonlinear Discriminant Analysis (NDA), we first pretrain a DNN using stochastic optimization, partly supervised and unsup...

Constructing a recognition system based on raw measurements for different objects usually requires expert knowledge of domain specific data preprocessing, feature extraction, and classifier design. We seek to simplify this process in a way that can be applied without any knowledge about the data domain and the specific properties of different class...

Die Erfindung betrifft ein Verfahren und eine Vorrichtung zur Detektion einer zweidimensionalen Darstellung des Gesichtes einer Person, derart, dass mit einer digitalen Kamera ein echtes (dreidimensionales) Gesicht oder eine zweidimensionale Darstellung eines Gesichtes aufgenommen wird, dabei optional eine spezielle Beleuchtungseinrichtung zum Eins...

Security is an essential requirement for casinos. Besides protection from fraud and vandalism, casinos are requested to control addicted gambling. This article looks at the German experience of implementing facial recognition to prevent gambling addicts from entering casino premises.

Face recognition systems attract much attention, scientifically, commercially, and ever again in the media. However, today's face recognition technology is far from being perfect. Face recognition also has to compete against other biometric methods. This paper has two parts: It first analyses the need for facial recognition systems from the point o...

Face recognition systems attract much attention, scientifically, commercially, and ever again in the media. However, today's face recognition technology is far from being perfect. Face recognition also has to compete against other biometric methods. This paper has two parts: It first analyses the need for facial recognition systems from the point o...

Vorrichtung zur Überwachung des Raumes eines Abschnittes eines Gleises in einem Bahnhof in Bezug auf die Anwesenheit von Objekten, die dieser im Normalzustand nicht aufweist, umfassend zwei über dem Gleisbereich angebrachte und längs des Gleises in der Aufsicht annähernd gegeneinander gerichtete Kameraeinheiten (1, 2) mit jeweils mindestens zwei a...

Hochauflösende bildgebende Sensoren (Kameras) sind prinzipiell geeignet das Vorhandensein von Personen, d.h. von menschlichen Gestalten, in einer Szene in ein elektronisch auswertbares Signal abzubilden. Die Defizite technischer Bildsensoren nach heutigem Stand der Technik können in absehbarer Zeit überwunden werden und einzelne Leistungsdaten von...

A method and device for automatically taking pictures of faces has a device for aligning and setting a camera to automatically take pictures of human faces, especially in connection with the electronic further processing (storage, identification, etc.) of face images. The face of a person is detected with the help of an opto-electronic image proces...

A system for recognizing authorization of a person to use a vehicle includes an optical-electronic image processor arranged to identify an authorized user of the vehicle by recognizing a distinctive physical feature of the user. The image processor provides an output signal which controls locking functions relating to operation of the vehicle based...

Eine zuverlässige und kostengünstige Alternative zu bisher bekannten Mehrlinien-Palettiersystemen wird seit kurzem bei der in lbbenbüren ansässigen Stärkefabrik Crespel & Deiters eingesetzt. Dreh- und Angelpunkt dieser Anlage ist ein neuartiges Produkterkennungssystem.

Automated processing of facial images has become a serious market for both hard-and software products. For the commercial success of face recognition systems it is most crucial that the process of face image capturing is very convenient for the people exposed to such systems. As a consequence, the whole imaging setup has to be carefully designed fo...

Automated processing of facial images has become a serious market for both hard- and software products. For the commercial success of face recognition systems it is most crucial that the process of face image capturing is very convenient for the people exposed to such systems. As a consequence, the whole imaging setup has to be carefully designed...

Fahrerloses Fahren im Nahverkehr wird in naher Zukunft große Bedeutung erhalten. Im Rahmen von Bemühungen den Nahverkehr attraktiver und kosteneffizienter zu machen, ist die Einführung von vollautomatischen fahrerlosen U-Bahn-Zügen ein technisch machbarer und wirtschaftlich notwendiger Schritt.

Abschlußbericht, 01.01.1995-31.12.1996, Sehsysteme für Fahrerassistenz im Automobil und Bewertung komplexer Umgebungen.

Verfahren zur Erkennung einer Benutzungsberechtigung eines Kraftfahrzeuges mit einer optoelektronischen Bildverarbeitungseinrichtung zur Identifizierung eines Fahrzeugbenutzers, wobei in Abhängigkeit mindestens eines Identifizierungssignals (A) der Bildverarbeitungseinrichtung (1) Sperrfunktionen (1 0) zur lnbetriebnahme des Kraftfahrzeuges gesteue...

Der vorliegende Beitrag behandelt das Thema Konturextraktion und Konturmodellierung. Konturen sind als Grenzen zwischen homogenen bzw. glatten Bildbereichen beschreibbar. Sie stellen symbolische Bildmerkmale dar und tragen einen großen Teil der visuellen Information. Ein Konturbild genügt einem Betrachter oftmals zum Verständnis der abgebildeten Sz...

This paper deals with the development and implementation of a purely visual obstacle detection framework for autonomous driving on motorways. Our activities are embedded in the SMART VEHICLE subproject of the ESPRIT project CLEOPATRA. The aim of SMART VEHICLE is the development of a visually controlled intelligent cruise control (ICC) for a prototy...

The presented work describes the realization of a machine vision approach for obstacle sensing in cars. Most of the techniques and approaches arose from the European PROMETHEUS program and from a variety research projects carried out at the institution of the author. The main topics include symmetry measuring, adaptive object recognition with neura...

We currently work on the implementation of an integrated prototype system that provides "all around" automatic visual obstacle sensing for a Daimler-Benz test car. Most of the machine vision techniques being used have been developed within the European PROMETHEUS programme and a number of other research projects carried out by the authors and other...

We currently work on the implementation of an integrated prototype system that provides "all around" automatic visual obstacle sensing for a Daimler-Benz test car. Most of the machine vision techniques being used have been developed within the European PROMETHEUS program and a number of other research projects carried out by the authors and other a...

We present two methods for detecting mirror symmetry in images, one based directly on the intensity values and another one based on a discrete representation of local orientation. A symmetry finder has been developed which uses the intensity-based method to search an image for compact regions which display some degree of mirror symmetry due to inte...

Die vorliegende Arbeit beschäftigt sich mit Problemen der visuellen Hinderniserkennung im Forschungsgebiet Rechnersehen für Automobile. Es werden Verfahren bzw. Prinzipien des Fachgebiets Computer–Vision angewendet, bewertet und erforderlichenfalls neu entwickelt, wobei stets das Anwendungsproblem eines video-gesteuerten Automobils im Vordergrund s...

We present two methods for detecting symmetry in images, one based directly on the intensity values and another one based on a discrete representation of local orientation. A symmetry finder has been developed which uses the intensity-based method to search an image for compact regions which display some degree of mirror symmetry due to intensity s...

CARTRACK is a computer vision system that can reliably detect, track, and measure vehicle rears in images from a video camera in a following car. The system exploits the symmetry property typical for the rear of most vehicles on normal roads. The authors present two novel methods for detecting mirror symmetry in images, one based directly on the in...

For image contours approximated by conic curve segments. The authors examine how to detect efficiently, given geometric relationships between conic segments. They present a method that can serve as a general tool for model-driven matching of conic curve segments

The authors have developed a vision system for automatic car following and object classification. The CARTRACK system can reliably detect, track, and identify the back of an automobile in a dynamic image taken from a following car in the same lane or in a neighbor lane. The detection and tracking system exploits the symmetry property of the rear vi...

Topographic mappings are neigbourhood preserving transformations between twodimensional data structures. Mappings of this type are a general means of information processing in the vertebrate visual system. In this paper we present an application of a special topographic mapping termed the inverse perspective mapping for the computation of stereo an...

Topographic mappings are neigbourhood preserving transformations between twodimensional data structures. Mappings of this type are a general means of information processing in the vertebrate visual system. In this paper we present an application of a special topographic mapping termed the inverse perspective mapping for the computation of stereo an...

We have developed a stereo obstacle detection system for automatically guided vehicles (AGVs) that operate on flat (factory) floors . Our system does not attempt to visually reconstruct the 3D environment but simply tries to answer the question 'Do I see anything else than the floor on my way ?'. A novel approach to stereo image processing is prese...

Topographic mappings are neigborhood preserving transformations between two-dimensional data structures. Mappings of this type are a general means of information processing in the vertebrate visual system. In this paper, we present an application of a special topographic mapping , termed the inverse perspective mapping, for the processing of stereo...

Bei einem Verfahren zum Führen von ohne Fahrer frei beweglichen Fahrzeugen auf einer Fahrbahn mittels am Fahrzeug vorhandener optischer Sensoren, die eine längs der Fahrbahn angeordnete unterbrochene Markierung aufnehmen und mit im Fahrzeug gespeicherten Führungsdaten vergleichen, wobei bei Abweichungen von den Führungsdaten Steuerbefehle zur Führu...

Jedes sehende Lebewesen orientiert sich problemlos in unbekannter Umgebung, selbst Insekten durchfliegen zügig kompliziertes Terrain und weichen sogar beweglichen Hindernissen mühelos aus. Ein mit einer Videokamera versehener Roboter tut sich bereits schwer, einfache Werkstücke auf einem Fließband zu erkennen. Durch die Erforschung neuronaler Archi...

The problem of describing the structural information that is contained in an image is crucial in computer vision. Methods in use are often based on edge detection and are typically designed for a limited application domain. It is widely believed that any general approach to low-level image description is bound to have two fundamental features: (1)...