Refine
Year of publication
- 2021 (82) (remove)
Document Type
- Conference Proceeding (46)
- Article (reviewed) (24)
- Letter to Editor (4)
- Book (3)
- Part of a Book (2)
- Doctoral Thesis (2)
- Report (1)
Conference Type
- Konferenzartikel (41)
- Konferenz-Abstract (5)
Language
- English (82) (remove)
Is part of the Bibliography
- yes (82)
Keywords
- Blockchain (3)
- Generative Adversarial Network (3)
- Götz von Berlichingen (3)
- neural networks (3)
- printed electronics (3)
- Machine Learning (2)
- Predictive maintenance (2)
- Stability (2)
- ancient Capua leg (2)
- artificial intelligence (2)
Institute
- Fakultät Elektrotechnik, Medizintechnik und Informatik (EMI) (ab 04/2019) (82) (remove)
Open Access
- Open Access (46)
- Closed Access (35)
- Bronze (1)
- Closed (1)
- Diamond (1)
- Grün (1)
Time-Sensitive Networking (TSN) is the most promising time-deterministic wired communication approach for industrial applications. To extend TSN to "IEEE 802.11" wireless networks two challenging problems must be solved: synchronization and scheduling. This paper is focused on the first one. Even though a few solutions already meet the required synchronization accuracies, they are built on expensive hardware that is not suited for mass market products. While next Wi-Fi generation might support the required functionalities, this paper proposes a novel method that makes possible high-precision wireless synchronization using commercial low-cost components. With the proposed solution, a standard deviation of synchronization error of less than 500 ns can be achieved for many use cases and system loads on both CPU and network. This performance is comparable to modern wired real-time field busses, which makes the developed method a significant contribution for the extension of the TSN protocol to the wireless domain.
Multiple Object Tracking (MOT) is a long-standing task in computer vision. Current approaches based on the tracking by detection paradigm either require some sort of domain knowledge or supervision to associate data correctly into tracks. In this work, we present a self-supervised multiple object tracking approach based on visual features and minimum cost lifted multicuts. Our method is based on straight-forward spatio-temporal cues that can be extracted from neighboring frames in an image sequences without supervision. Clustering based on these cues enables us to learn the required appearance invariances for the tracking task at hand and train an AutoEncoder to generate suitable latent representations. Thus, the resulting latent representations can serve as robust appearance cues for tracking even over large temporal distances where no reliable spatio-temporal features can be extracted. We show that, despite being trained without using the provided annotations, our model provides competitive results on the challenging MOT Benchmark for pedestrian tracking.
An Empirical Study of Explainable AI Techniques on Deep Learning Models For Time Series Tasks
(2021)
Decision explanations of machine learning black-box models are often generated by applying Explainable AI (XAI) techniques. However, many proposed XAI methods produce unverified outputs. Evaluation and verification are usually achieved with a visual interpretation by humans on individual images or text. In this preregistration, we propose an empirical study and benchmark framework to apply attribution methods for neural networks developed for images and text data on time series. We present a methodology to automatically evaluate and rank attribution techniques on time series using perturbation methods to identify reliable approaches.
A Hybrid Optoelectronic Sensor Platform with an Integrated Solution‐Processed Organic Photodiode
(2021)
Hybrid systems, unifying printed electronics with silicon‐based technology, can be seen as a driving force for future sensor development. Especially interesting are sensing elements based on printed devices in combination with silicon‐based high‐performance electronics for data acquisition and communication. In this work, a hybrid system integrating a solution‐processed organic photodiode in a silicon‐based system environment, which enables flexible device measurement and application‐driven development, is presented. For performance evaluation of the integrated organic photodiode, the measurements are compared to a silicon‐based counterpart. Therefore, the steady state response of the hybrid system is presented. Promising application scenarios are described, where a solution‐processed organic photodiode is fully integrated in a silicon system.
To demonstrate how deep learning can be applied to industrial applications with limited training data, deep learning methodologies are used in three different applications. In this paper, we perform unsupervised deep learning utilizing variational autoencoders and demonstrate that federated learning is a communication efficient concept for machine learning that protects data privacy. As an example, variational autoencoders are utilized to cluster and visualize data from a microelectromechanical systems foundry. Federated learning is used in a predictive maintenance scenario using the C-MAPSS dataset.
Dementia is a clinical diagnosis reflecting many possible underlying pathologies, for example, vascular dementia and neurodegenerative disorders such as frontotemporal dementia, Lewy body-type disorder or Alzheimer’s disease (AD). The breakthrough of 99mtechnetium-labelled perfusion tracers in the 1990s resulted in many SPECT studies of flow changes in AD. In the first decade of 2000, the role of perfusion SPECT was shifted from diagnosis towards differential diagnosis, parallel to the growing attention for diagnosing early stages of dementia. Previously a diagnosis based largely on a process of exclusion, new guidelines have emerged increasingly employing positive criteria to establish the diagnosis, including neuroimaging biomarkers. Nowadays, FDG PET has largely limited the role of perfusion SPECT, although it is still considered a valuable and cost-effective alternative when PET is not available.
Investigation of the Angle Dependency of Self-Calibration in Multiple-Input-Multiple-Output Radars
(2021)
Multiple-Input-Multiple-Output (MIMO) is a key technology in improving the angular resolution (spatial resolution) of radars. In MIMO radars the amplitude and phase errors in antenna elements lead to increase in the sidelobe level and a misalignment of the mainlobe. As the result the performance of the antenna channels will be affected. Firstly, this paper presents analysis of effect of the amplitude and phase errors on angular spectrum using Monte-Carlo simulations. Then, the results are compared with performed measurements. Finally, the error correction with a self-calibration method is proposed and its angle dependency is evaluated. It is shown that the values of the errors change with an incident angle, which leads to a required angle-dependent calibration.
PET and SPECT in Psychiatry
(2021)
This book provides a comprehensive overview of the use of PET and SPECT in the classic psychiatric disorders such as depression, bipolar disorder, anxiety disorders, and schizophrenia. In addition, it discusses the application of these functional neuroimaging techniques in a variety of other conditions, including sleep disorders, eating disorders, autism, and chronic fatigue syndrome. The new edition has been extensively revised and updated to reflect the latest advances and results in nuclear imaging within the field. Most chapters are written jointly by a clinical psychiatrist and a nuclear medicine expert to ensure a multidisciplinary approach. This state of the art compendium will be of value for all who have an interest in the field of neuroscience, from psychiatrists and radiologists/nuclear medicine specialists to interested general practitioners and cognitive psychologists. Companion volumes on the use of PET and SPECT in neurology and for the imaging of neurobiological systems complete a trilogy.