Informatik
In order to generate a machine learning algorithm (MLA) that can support ophthalmologists with the diagnosis of glaucoma, a carefully selected dataset that is based on clinically confirmed glaucoma patients as well as borderline cases (e.g., patients with suspected glaucoma) is required. The clinical annotation of datasets is usually performed at the expense of the data volume, which results in poorer algorithm performance. This study aimed to evaluate the application of an MLA for the automated classification of physiological optic discs (PODs), glaucomatous optic discs (GODs), and glaucoma-suspected optic discs (GSODs). Annotation of the data to the three groups was based on the diagnosis made in clinical practice by a glaucoma specialist. Color fundus photographs and 14 types of metadata (including visual field testing, retinal nerve fiber layer thickness, and cup–disc ratio) of 1168 eyes from 584 patients (POD = 321, GOD = 336, GSOD = 310) were used for the study. Machine learning (ML) was performed in the first step with the color fundus photographs only and in the second step with the images and metadata. Sensitivity, specificity, and accuracy of the classification of GSOD vs. GOD and POD vs. GOD were evaluated. Classification of GOD vs. GSOD and GOD vs. POD performed in the first step had AUCs of 0.84 and 0.88, respectively. By combining the images and metadata, the AUCs increased to 0.92 and 0.99, respectively. By combining images and metadata, excellent performance of the MLA can be achieved despite having only a small amount of data, thus supporting ophthalmologists with glaucoma diagnosis.
Biometric fingerprint identification hinges on the reliability of its sensors; however, calibrating and standardizing these sensors poses significant challenges, particularly in regards to repeatability and data diversity. To tackle these issues, we propose methodologies for fabricating synthetic 3D fingerprint targets, or phantoms, that closely emulate real human fingerprints. These phantoms enable the precise evaluation and validation of fingerprint sensors under controlled and repeatable conditions. Our research employs laser engraving, 3D printing, and CNC machining techniques, utilizing different materials. We assess the phantoms’ fidelity to synthetic fingerprint patterns, intra-class variability, and interoperability across different manufacturing methods. The findings demonstrate that a combination of laser engraving or CNC machining with silicone casting produces finger-like phantoms with high accuracy and consistency for rolled fingerprint recordings. For slap recordings, direct laser engraving of flat silicone targets excels, and in the contactless fingerprint sensor setting, 3D printing and silicone filling provide the most favorable attributes. Our work enables a comprehensive, method-independent comparison of various fabrication methodologies, offering a unique perspective on the strengths and weaknesses of each approach. This facilitates a broader understanding of fingerprint recognition system validation and performance assessment.
We address the need for a large-scale database of children’s faces by using generative adversarial networks (GANs) and face-age progression (FAP) models to synthesize a realistic dataset referred to as “HDA-SynChildFaces”. Hence, we proposed a processing pipeline that initially utilizes StyleGAN3 to sample adult subjects, which is subsequently progressed to children of varying ages using InterFaceGAN. Intra-subject variations, such as facial expression and pose, are created by further manipulating the subjects in their latent space. Additionally, this pipeline allows the even distribution of the races of subjects, allowing the generation of a balanced and fair dataset with respect to race distribution. The resulting HDA-SynChildFaces consists of 1,652 subjects and 188,328 images, each subject being present at various ages and with many different intra-subject variations. We then evaluated the performance of various facial recognition systems on the generated database and compared the results of adults and children at different ages. The study reveals that children consistently perform worse than adults on all tested systems and that the degradation in performance is proportional to age. Additionally, our study uncovers some biases in the recognition systems, with Asian and black subjects and females performing worse than white and Latino-Hispanic subjects and males.
Solar phase scintillation and solar amplitude scintillation are fundamentally important in deep space mission operations for designing a communication system capable of transmitting signals when the signal path is close to the Sun. The ESA’s BepiColombo measurement data were analyzed in a previous paper in terms of the power spectral density of the solar phase scintillation, also with a comparison with Woo’s solar phase scintillation theory, when X-band and Ka-band signals propagate close to the Sun with a small Sun-Earth-Probe (SEP) angle during the superior solar conjunction campaign in March 2021 in its cruise phase to Mercury. In this paper the solar amplitude scintillation is analyzed both by calculating the power spectral density and the scintillation index. The results of scintillation index, derived from these measurement data, fit the NASA JPL’s scintillation index model.
The relevance of Machine Intelligence, a.k.a. Artificial Intelligence (AI), is undisputed at the present time. This is not only due to AI successes in research but, more prominently, its use in day-to-day practice. In 2014, we started a series of annual workshops at the Leibniz Zentrum für Informatik, Schloss Dagstuhl, Germany, initially focussing on Corporate Semantic Web and later widening the scope to Applied Machine Intelligence. This article presents a number of AI applications from various application domains, including medicine, industrial manufacturing and the insurance sector. Best practices, current trends, possibilities and limitations of new AI approaches for developing AI applications are also presented. Focus is put on the areas of natural language processing, ontologies and machine learning. The article concludes with a summary and outlook.
Signals and images with discontinuities appear in many problems in such diverse areas as biology, medicine, mechanics and electrical engineering. The concrete data are often discrete, indirect and noisy measurements of some quantities describing the signal under consideration. A frequent task is to find the segments of the signal or image which corresponds to finding the discontinuities or jumps in the data. Methods based on minimizing the piecewise constant Mumford–Shah functional—whose discretized version is known as Potts energy—are advantageous in this scenario, in particular, in connection with segmentation. However, due to their non-convexity, minimization of such energies is challenging. In this paper, we propose a new iterative minimization strategy for the multivariate Potts energy dealing with indirect, noisy measurements. We provide a convergence analysis and underpin our findings with numerical experiments.
Machine intelligence, a.k.a. artificial intelligence (AI) is one of the most prominent and relevant technologies today. It is in everyday use in the form of AI applications and has a strong impact on society. This article presents selected results of the 2020 Dagstuhl workshop on applied machine intelligence. Selected AI applications in various domains, namely culture, education, and industrial manufacturing are presented. Current trends, best practices, and recommendations regarding AI methodology and technology are explained. The focus is on ontologies (knowledge-based AI) and machine learning.
The awareness of emerging trends is essential for strategic decision making because technological trends can affect a firm’s competitiveness and market position. The rise of artificial intelligence methods allows gathering new insights and may support these decision-making processes. However, it is essential to keep the human in the loop of these complex analytical tasks, which, often lack an appropriate interaction design. Including special interactive designs for technology and innovation management is therefore essential for successfully analyzing emerging trends and using this information for strategic decision making. A combination of information visualization, trend mining and interaction design can support human users to explore, detect, and identify such trends. This paper enhances and extends a previously published first approach for integrating, enriching, mining, analyzing, identifying, and visualizing emerging trends for technology and innovation management. We introduce a novel interaction design by investigating the main ideas from technology and innovation management and enable a more appropriate interaction approach for technology foresight and innovation detection.
When NoSQL database systems are used in an agile software development setting, data model changes occur frequently and thus, data is routinely stored in different versions. The management of versioned data leads to an overhead potentially impeding the software development. Several data migration strategies exist that handle legacy data differently during data accesses, each of which can be characterized by certain advantages and disadvantages. Depending on the requirements for the software application, we evaluate and compare different migration strategies through metrics like migration costs and latency as well as precision and recall. Ideally, exactly that strategy should be selected whose characteristics fulfill service-level agreements and match the migration scenario, which depends on the query workload and the changes in the data model which imply an evolution of the database schema. In this paper, we present a methodology of self-adapting data migration, which automatically adjusts migration strategies and their parameters with respect to the migration scenario and service-level agreements, thereby contributing to the self-management of database systems and supporting agile development.
Smart factories are complex; with the increased complexity of employed cyber-physical systems, the complexity evolves further. Cyber-physical systems produce high amounts of data that are hard to capture and challenging to analyze. Real-time recording of all data is not possible due to limited network capabilities. Limited network capabilities are the reason for a chain of faults introduced via active surveillance during fault diagnosis. These introduced faults may slow down production or lead to an outage of the production line. Here, we present a novel approach to automatically select production-relevant shop floor parameters to decrease the number of surveyed variables and, at the same time, maintain quality in fault diagnosis without overloading the network. We were able to achieve higher throughput, mitigate communication losses and prevent the disruption of factory instructions. Our approach uses an autoencoder ensemble via minority voting to differentiate between normal—always on—variables and production variables that may yield a higher entropy. Our approach has been tested in a production-equal smart factory and was cross-validated by a domain expert.
Cyber-physical systems become more complex, therewith production lines become more complex in the smart factory. Every employed system produces high amounts of data with unknown dependencies and relationships, making incident reasoning difficult. Context-aware fault diagnosis can unveil such relationships on different levels. A fault diagnosis application becomes context-aware when the current production situation is used in the reasoning process. We have already published TAOISM, a visual analytics model defining the context-aware fault diagnosis process for the Industry 4.0 domain. In this article, we propose the Flourish dashboard for context-aware fault diagnosis. The eponymous visualization Flourish is a first implementation of a context-displaying visualization for context-aware fault diagnosis in an Industry 4.0 setting. We conducted a questionnaire and interview-based bilingual evaluation with two user groups based on contextual faults recorded in a production-equal smart factory. Both groups provided qualitative feedback after using the Flourish dashboard. We positively evaluate the Flourish dashboard as an essential part of the context-aware fault diagnosis and discuss our findings, open gaps, and future research directions.
Within the last few decades, the need for subject authentication has grown steadily, and biometric recognition technology has been established as a reliable alternative to passwords and tokens, offering automatic decisions. However, as unsupervised processes, biometric systems are vulnerable to presentation attacks targeting the capture devices, where presentation attack instruments (PAI) instead of bona fide characteristics are presented. Due to the capture devices being exposed to the public, any person could potentially execute such attacks. In this work, a fingerprint capture device based on thin film transistor (TFT) technology has been modified to additionally acquire the impedances of the presented fingers. Since the conductance of human skin differs from artificial PAIs, those impedance values were used to train a presentation attack detection (PAD) algorithm. Based on a dataset comprising 42 different PAI species, the results showed remarkable performance in detecting most attack presentations with an APCER = 2.89% in a user-friendly scenario specified by a BPCER = 0.2%. However, additional experiments utilising unknown attacks revealed a weakness towards particular PAI species.
Mobile Contactless Fingerprint Recognition: Implementation, Performance and Usability Aspects
(2022)
This work presents an automated contactless fingerprint recognition system for smartphones.We provide a comprehensive description of the entire recognition pipeline and discuss important requirements for a fully automated capturing system. In addition, our implementation
is made publicly available for research purposes. During a database acquisition, a total number of 1360 contactless and contact-based samples of 29 subjects are captured in two different environmental
situations. Experiments on the acquired database show a comparable performance of our contactless scheme and the contact-based baseline scheme under constrained environmental influences. A comparative usability study on both capturing device types indicates that the majority of subjects prefer the contactless capturing method. Based on our experimental results, we analyze the impact of the current COVID-19 pandemic on fingerprint recognition systems. Finally, implementation aspects of contactless fingerprint recognition are summarized.
The growing scope, scale, and number of biometric deployments around the world emphasise the need for research into technologies
facilitating efficient and reliable biometric identification queries. This work presents a method of indexing biometric databases,
which relies on signal-level fusion of facial images (morphing) to create a multi-stage data-structure and retrieval protocol. By
successively pre-filtering the list of potential candidate identities, the proposed method makes it possible to reduce the necessary
number of biometric template comparisons to complete a biometric identification transaction. The proposed method is extensively
evaluated on publicly available databases using open-source and commercial off-the-shelf recognition systems. The results show
that using the proposed method, the computational workload can be reduced down to around 30%, while the biometric performance
of a baseline exhaustive search-based retrieval is fully maintained, both in closed-set and open-set identification scenarios.
As an indispensable component of today’s world economy and an increasing success factor in production and other processes, as well as products, software needs to handle a growing number of specific requirements and influencing factors that are driven by globalization. Two common success factors in the domain of Software Systems Engineering are standardized software development processes and process-supported toolchains. Development processes should be formally integrated with toolchains. The sequence and the results of toolchains must also be validated with the specifications of the development process on several levels. The outcome of a conceptual deductive analysis is that there is neither a formal general mapping nor a generally accepted validation mechanism for the challenges that such an integrated concept faces. To close this research gap, this paper focuses on the core issue of the integration of development processes and toolchains in order to create benefits for modeling and automatization in the domain of systems engineering. Therefore, it describes a self-developed integration approach related to the recently introduced prototypical technical implementation TOPWATER. A unified metamodel specifies how processes and toolchains are linked by a general mapping mechanism that considers test options for the structural, content, and semantic levels.
In Flutter gibt es diverse Ansätze und Lösungsmöglichkeiten, den Zustand einer mobilen Anwendung zu verwalten. Flutter ist ein beliebtes Cross-Plattform-Framework, mit dem sich Anwendungen für iOS, Android, Web, Windows, Linux und macOS erstellen lassen, die den gleichen Quelltext verwenden. Damit sollen in der Entwicklung Aufwände gespart werden können und die Komplexität reduziert werden. Flutter baut auf eine deklarative Benutzeroberfläche, welche anhand eines Zustands erstellt wird. Die Verwaltung dieses Zustands ist entscheidend für die Architektur und die Funktion einer Anwendung.
In dieser Ausarbeitung werden verschieden bereits etablierte Ansätze zur Verwaltung des Zustands einer Flutter-Anwendung dargestellt und untersucht. Dabei ist das Ziel, herauszufinden, welcher Ansatz am besten zum Verwalten des Zustands einer Flutter-Anwendung geeignet ist. Dafür wird eine Evaluation für die Zustandsverwaltungssysteme setState, InheritedWidget, BLoC, Provider, Riverpod, Redux und MobX durchgeführt. Grundlage dieser Evaluation ist die Entwicklung einer Beispielanwendung für jedes Zustandsverwaltungssystem und die Bewertung dieser anhand von qualitative und quantitativen Bewertungskriterien, die anhand der Anforderungen an Zustandsverwaltungssysteme definiert werden.
Das Ergebnis der Arbeit stellt neben den Resultaten der Evaluation auch eine Empfehlung dar, welches Zustandsverwaltungssystem für welchen Anwendungsfall am besten genutzt werden sollte.
AutoML solutions provide the functionality to find optimized machine learning solutions for a given dataset without manual network configuration or hyperparameter optimisation.
Auto-PyTorch and AutoCVE are two such solutions which this paper will introduce regarding their functionality and compare based on their performance, usability and their features.
Automated machine learning (AutoML) supports the development of machine learning systems by automating tasks that traditionally require manual, time-consuming labor and extensive expertise in the field.
This paper presents two AutoML systems, one open-source (EvalML) and one commercial solution (Azure AutoML), and evaluates their range of functionality as well as practical usability.
The results show that commercial tools provide better usability for beginners, while open-source alternatives tend to stand out with greater configurability.
Die vorliegende Arbeit vergleicht sowohl die technischen Möglichkeiten als auch die Benchmarking Performance von AutoGluon und TPOT miteinander. Die Experimente zeigen, dass AutoGluon strukturierte Rohdaten besser verarbeitet. TPOT kann jedoch nur mit numerischen Features arbeiten. Daher sollen alle kategoriellen Features in den Rohdaten vorher in numerische konvertiert werden, bevor auf den Daten Modelle mit TPOT angepasst werden können. TPOT eignet sich für Klassifikations- und Regressionstasks. AutoGluon kann zusätzlich für Objektklassifikation und -Erkennung auf Bilddaten sowie für NLP Tasks auf Textdaten verwendet werden. Bei dem Regressionsexperiment erzielte das TPOT-Modell einen kleineren RMSE-Wert als das AutoGluon-Modell. Der F1-Score von AutoGluon war bei dem Klassifikationsexperiment höher als der des TPOT-Modells.
We compare the open-source library ATM with the fully managed
Amazon Web Services solution Sagemaker Autopilot. Depending of prior
knowledge and experience there are found individual advantages. Two Datasets
are tested representing a classification and a regression task. During classification,
Autopilot achieves an F1 score of 0.971 whereas the average F1 score of
ATM is 0.916. Due to the limited available budget settings in Autopilot, this
comparison has marginal value. The results of the regression dataset cannot be
compared due to the missing functionality of ATM to calculate regression problems.
The motivation and need for Automated Machine Learning is discussed
and benefits are given, explained with the two practical examples. We expect a
great demand of Automated Machine Learning solutions with growing machine
learning problems and a limited number of experts in the field. The available
models and algorithms within the solutions presented here are discussed and a
missing ongoing development and support for ATM is found whereas Autopilot
is provided with regular updates. The experiments are limited to the local usage
of ATM and focused on the GUI approach with Autopilot. There is an additional
option to deploy ATM in a distributed and scalable manner and to use Autopilot
only with Jupyter Notebooks or as Python SDK. These options are to be evaluated
in further research.