A5008053208- Remote Sensing and LiDAR Applications
- Remote Sensing in Agriculture
- Speech and dialogue systems
- Context-Aware Activity Recognition Systems
- Species Distribution and Climate Change
- Housing Market and Economics
- Plant Water Relations and Carbon Dynamics
- Land Use and Ecosystem Services
- Cloud Computing and Resource Management
- Urban Planning and Valuation
- Image Processing and 3D Reconstruction
- Monetary Policy and Economic Impact
- IoT and Edge/Fog Computing
- Natural Language Processing Techniques
- Economic Growth and Productivity
- Time Series Analysis and Forecasting
- Image Retrieval and Classification Techniques
- AI in cancer detection
- Semantic Web and Ontologies
- Radiology practices and education
- Climate Change and Health Impacts
- Sustainable Building Design and Assessment
- Multimedia Communication and Technology
- Digital Radiography and Breast Imaging
- Scientific Computing and Data Management
Technische Universität Berlin
2016-2024
Helmholtz Centre for Environmental Research
2024
Otto-von-Guericke University Magdeburg
2014-2023
Thüringer Universitäts- und Landesbibliothek
2023
University Hospital Regensburg
2023
GeoInformation (United Kingdom)
2022
German Research Centre for Artificial Intelligence
2010-2015
Halle Institute for Economic Research
2014-2015
Technical University of Darmstadt
2015
Düsseldorf University Hospital
2014
Abstract. Airborne and spaceborne platforms are the primary data sources for large-scale forest mapping, but visual interpretation individual species determination is labor-intensive. Hence, various studies focusing on forests have investigated benefits of multiple sensors automated tree classification. However, transferable deep learning approaches applications still lacking. This gap motivated us to create a novel dataset classification in central Europe based multi-sensor from aerial,...
The impacts of global change, including extreme heat and water scarcity, are threatening an ever-growing urban world population. Evapotranspiration (ET) mitigates the island, reducing effect waves. It can also be used as a proxy for vegetation use, making it crucial tool to plan resilient green cities. To optimize trade-off between greening security, reliable up-to-date maps ET cities urgently needed. Despite its importance, few studies have mapped accurately entire city in high spatial...
This paper is the first to propose valid inference tools, based on self-normalization, in time series expected shortfall regressions. In doing so, we a novel two-step estimator for regressions which convex optimization both steps (rendering computation easy) and it only requires minimization of quantile losses squared error (methods are implemented every standard statistical computing package). As corollary, also derive self-normalized tools Extant methods, bootstrap or direct estimation...
Tree species maps derived from satellite imagery increasingly support forest administrations and nature conservation authorities with large-scale up-to-date information. However, many are often excluded or aggregated in classification tasks due to a limited knowledge of the most suitable predictors. Our study aims gain better understanding optical polarimetric traits for tree mapping by examining Sentinel-1 Sentinel-2 time series 61 temperate Europe. For selection 32 optical, structural...
Abstract. Airborne and spaceborne platforms are the primary data sources for large-scale forest mapping, but visual interpretation individual species determination is labour-intensive. Hence, various studies focusing on forests have investigated benefits of multiple sensors automated tree classification. However, transferable deep learning approaches applications still lacking. This gap motivated us to create a novel dataset classification in Central Europe based multi-sensor from aerial,...
The collection of a high number pixel-based labeled training samples for tree species identification is time consuming and costly in operational forestry applications. To address this problem, paper we investigate the effectiveness explanation methods deep neural networks performing weakly supervised semantic segmentation using only image-level labels. Specifically, consider four methods: i) class activation maps (CAM); ii) gradient-based CAM; iii) pixel correlation module; iv)...
With RadSpeech, we aim to build the next generation of intelligent, scalable, and user-friendly semantic search interfaces for medical imaging domain, based on technologies. Ontology-based knowledge representation is used not only image contents, but also complex natural language understanding dialogue management process. This demo shows a speech-based annotation system radiology images focuses new effective way annotate regions with specific medical, structured, diagnosis while using speech...
We present two user interfaces: one multimodal dialogue system and task-based calendar which assist people with mild cognitive disabilities affecting their concentration memory. A new middleware based upon a open industrial standard—ISO/IEC 24752 Universal Remote Console (URC)—allows access to any network services or appliances as well devices for home entertainment household via abstract interfaces. This architecture promotes the concept of pluggable interfaces, that is, interface being...
The paper presents an analysis of the trade-offs participants different type between payment delay and liquidity requirement on basis synthetically generated data. generation synthetic transaction data set for a simple RTGS system is described calibrated using real world parameters. simulated various levels it shown that size in terms volume value will have optimal requirements, as delays they face each level be different. This indifference curves requirements.
Consumers' rapid adoption of new products is a key factor in firms' market success. To adequately address consumer requirements and thereby increase their technological products, firms need knowledge about consumers' predispositions toward those products. Some consumers are simply more ready to adopt than others, which leaves puzzled as how they should optimally This study proposes taxonomy, based on Relying trait theory data from 738 consumers, it identifies four patterns that differ...
Robust infrastructures for managing and accessing high volume data are an essential foundation unraveling complex spatiotemporal processes in the earth system sciences. Addressing multifaceted research questions demands from diverse sources; however, isolated solutions hinder effective collaboration knowledge advancement.We present a novel digital ecosystem FAIR time series management, deeply rooted contemporary software engineering developed at Helmholtz Centre Environmental Research (UFZ)...
We show that the first-order theory of Sturmian words over Presburger arithmetic is decidable. Using a general adder recognizing addition in Ostrowski numeration systems by Baranwal, Schaeffer and Shallit, we prove expansions single word are uniformly $\omega$-automatic, then deduce decidability class such structures. an implementation this decision algorithm called Pecan, automatically reprove classical theorems about seconds, able to obtain new results antisquares antipalindromes...