Dynamic Switching State Systems For Visual Tracking

This work addresses the problem of how to capture the dynamics of maneuvering objects for visual tracking. Towards this end, the perspective of recursive Bayesian filters and the perspective of deep learning approaches for state estimation are considered and their functional viewpoints are brought together.

In August 2022, Fraunhofer IOSB and IES of KIT held a joint workshop in a Schwarzwaldhaus near Triberg. Doctoral students presented research reports and discussed various topics like computer vision, optical metrology, network security, usage control, and machine learning. This book compiles the workshop's results and ideas, offering a comprehensive overview of the research program of IES and Fraunhofer IOSB.

In 2020 fand der jährliche Workshop des Faunhofer IOSB und the Lehrstuhls für interaktive Echtzeitsysteme statt. Vom 27. bis zum 31. Juli trugen die Doktorranden der beiden Institute über den Stand ihrer Forschung vor in Themen wie KI, maschinellen Lernen, computer vision, usage control, Metrologie vor. Die Ergebnisse dieser Vorträge sind in diesem Band als technische Berichte gesammelt. - In 2020, the annual joint workshop of the Fraunhofer IOSB and the Vision and Fusion Laboratory of the KIT was hosted at the IOSB in Karlsruhe. For a week from the 27th to the 31st July the doctoral students of both institutions presented extensive reports on the status of their research and discussed topics ranging from computer vision and optical metrology to network security, usage control and machine learning. The results and ideas presented at the workshop are collected in this book.

2021, the annual joint workshop of the Fraunhofer IOSB and KIT IES was hosted at the IOSB in Karlsruhe. For a week from the 2nd to the 6th July the doctoral students extensive reports on the status of their research. The results and ideas presented at the workshop are collected in this book in the form of detailed technical reports.

This work proposes a probabilistic extension to Bézier curves as a basis for effectively modeling stochastic processes with a bounded index set. The proposed stochastic process model is based on Mixture Density Networks and Bézier curves with Gaussian random variables as control points. A key advantage of this model is given by the ability to generate multi-mode predictions in a single inference step, thus avoiding the need for Monte Carlo simulation.

This book proposes a novel deep learning based detection method, focusing on vehicle detection in aerial imagery recorded in top view. The base detection framework is extended by two novel components to improve the detection accuracy by enhancing the contextual and semantical content of the employed feature representation. To reduce the inference time, a lightweight CNN architecture is proposed as base architecture and a novel module that restricts the search area is introduced.

The understanding and interpretation of complex 3D environments is a key challenge of autonomous driving. Lidar sensors and their recorded point clouds are particularly interesting for this challenge since they provide accurate 3D information about the environment. This work presents a multimodal approach based on deep learning for panoptic segmentation of 3D point clouds. It builds upon and combines the three key aspects multi view architecture, temporal feature fusion, and deep sensor fusion.

Configuring an anomaly-based Network Intrusion Detection System for cybersecurity of an industrial system in the absence of information on networking infrastructure and programmed deterministic industrial process is challenging. Within the research work, different self-learning frameworks to analyze passively captured network traces from PROFINET-based industrial system for protocol-based and process behavior-based anomaly detection are developed, and evaluated on a real-world industrial system.