share it - The Open Access and Research Data Repository of the University Libraries in Saxony-Anhalt
Not a member yet
91636 research outputs found
Sort by
MLUH-E-100_1, Falco tinnunculus canariensis (Koenig, AF, 1889), eggshell
No full textInventory No.: MLUH-E-100_1, Object: eggshell, Species: Falco tinnunculus canariensis (Koenig, AF, 1889), Preservation: complete preservation, Locality_loc.: Mirdouro Funchal Madeira, Locality today: Funchal, Country: PortugalCollector_leg.: Pater E. Schmitz, Date: 15/06/1906, Collection_coll.: M. Schönwetter, published in Handbuch der Oologie, Schönwetter, Max: Vol. I, p. 193, Identification by: M. Schönwetter, ex Collection: ex. Coll. Pater E. Schmit
Deep learning-based image optimization for low-field MRI
No full textPoint-of-care low-field MRI systems become increasingly interesting in modern medicine because of their small footprint and cost-effective profile. The portability of such devices poses new challenges, as they are often operated under uncontrolled and changing environmental conditions. The low magnetic field (50mT) and the high B0 field inhomogeneities, in combination with temperature fluctuations, and electromagnetic interferences culminate in a poor signal-to-noise ratio and severe image artifacts. The current state of research lacks suitable and robust correction methods that can be implemented directly into point-of-care devices using high-performance computing systems.
In this work, the combination of a novel, highly adaptable MR console and deep learning based image reconstructions is presented and evaluated to boost the imaging performance of point-of-care low-field MRI devices. The console shows comparable in vivo imaging results to state-of-the-art proprietary systems with additional features such as a fully transparent acquisition process, integration of auxiliary sensors and direct application of deep-learning based image reconstruction techniques.
By combining data driven image denoising and B0 field prediction with model-based image reconstruction, a physics-informed end-to-end reconstruction network is introduced and evaluated for low-field MRI.
The proposed method outperforms conventional reconstruction techniques based on the training and evaluation with simulated low-field MRI data. However, when the model is applied to real in vivo low-field MRI data, considerable challenges are identifiable. To address these challenges, a conventional regularization approach is combined with the estimation of regularization parameter maps by a neural network in an iterative approach in order to optimize image denoising. Thereby, the applicability of the neural network to real low-field MRI data could be improved. Finally, it could be shown that a model trained with supervision can be adapted to other data by applying a subsequent self-supervised refinement step. By integrating the methods and strategies presented in this work, low-field MRI image quality is successfully improved, enhancing its diagnostic value for future clinical applications.Niedrigfeld MRT Systeme für die Point-of-Care Anwendung erfreuen sich aufgrund ihres geringen Platzbedarfs und ihrer kostengünstigen Nutzung an zunehmendem Interesse in der modernen Medizin. Die Mobilität derartiger Systeme birgt jedoch neue Herausforderungen, da sie unter anderem bei veränderlichen und unkontrollierten Umgebungsbedingungen betrieben werden. Das niedrige Magnetfeld (50mT) und die hohen B0 Feldinhomogenitäten in Kombination mit Temperaturschwankungen und elektromagnetischen Interferenzen resultieren in einem schlechten Signal-zu-Rausch-Verhältnis und starken Bildartefakte.
Im aktuellen Stand der Forschung fehlen passende und robuste Korrekturverfahren, die mit leistungsfähigen Computersystemen direkt in Point-of-Care-Geräte implementiert werden können.
In dieser Arbeit, wird eine neuartige, anpassbare MR Konsole in Kombination mit Deep Learning-basierter Bildrekonstruktion vorgestellt und evaluiert, um die Bildgebungsleistung von Point-of-Care-Niedrigfeldsystemen zu optimieren. Die Konsole erlaubt In-vivo Bildgebung, die mit proprietären Systemen auf dem neuesten Stand der Technik vergleichbar ist, und ermöglicht darüber hinaus zusätzliche Funktionen wie einen vollständig transparenten Akquisitionsprozess, die Integration zusätzlicher Sensoren sowie die direkte Anwendung Deep-Learning-basierter Bildrekonstruktionstechniken.
Durch die Kombination von datengetriebener Rauschunterdrückung und B0 Feldschätzung mit modellbasierter Bildrekonstruktion wird ein physikalisch-informiertes End-to-End- Rekonstruktionsnetzwerk für die Niedrigfeld-MRT vorgestellt und evaluiert. Die vorgeschlagene Methode übertrifft konventionelle Rekonstruktionsverfahren basierend auf Training und Evaluation mit simulierten Niedrigfeld-MRT-Daten.
Bei der Anwendung auf reale in vivo Niedrigfeld-MRT-Daten, wurden jedoch erhebliche Herausforderungen identifiziert. Um diese Herausforderungen zu adressieren, wird ein alternativer Ansatz untersucht, der konventionelle Regularisierung mit der Schätzung von Regularisierungsparameterkarten durch ein neuronales Netzwerk in einem iterativen Ansatz kombiniert, um die Rauschunterdrückung zu optimieren. Dadurch konnte die Übertragbarkeit des neuronalen Netzes auf reale Niedrigfeld-MRT-Daten verbessert werden. Schließlich konnte gezeigt werden, dass ein mit Supervision trainiertes Modell durch einen nachfolgenden selbst überwachten Verfeinerungsschritt an andere Daten angepasst werden kann. Durch die Integration der in dieser Arbeit vorgestellten Methoden und Strategien wird die Bildqualität der Low- Field-MRT erfolgreich verbessert, wodurch ihr diagnostischer Wert für zukünftige klinische Anwendungen gesteigert wird.Literaturverzeichnis: Seite 129-14
Regenerative cartilage treatment for focal chondral defects in the knee : focus on marrow-stimulating and cell-based scaffold approaches
No full textFocal chondral defects of the knee are a common cause of pain and functional limitation in active individuals and may predispose to early degenerative joint changes. Given the limited regenerative capacity of hyaline cartilage, biologically based surgical strategies have emerged to promote tissue repair and restore joint function. This narrative review critically examines current treatment approaches that rely on autologous cell sources and scaffold-supported regeneration. Particular emphasis is placed on techniques that stimulate endogenous repair or support chondrocyte-based tissue restoration through the use of autologous biomaterial constructs. The influence of lesion morphology, joint biomechanics, and patient-specific variables on treatment selection is discussed in detail, focusing on the differences between tibiofemoral and patellofemoral involvement. Biologically driven approaches have shown promising mid- to long-term outcomes in selected patients, and are increasingly favoured over traditional methods in specific clinical scenarios. However, the literature remains limited by heterogeneity in study design, follow-up duration, and outcome measures. This review aims to provide an evidence-based, morphology-informed framework to support the clinical decision-making process in the management of knee cartilage defects
Investigations on the occurrence of West Nile virus, Usutu virus and Sindbis virus RNA in avian louse flies (Diptera: Hippoboscidae) collected in Germany (2016-2022)
No full textBackground:
As living vectors, arthropods play a crucial role in the transmission of viruses, bacteria and parasites. Previous research on virus transmission has focussed mainly on the roles of mosquitoes and ticks, while the potential importance of other blood-sucking arthropods such as louse flies (Hippoboscidae) has been somewhat neglected. The aim of this study was to detect viruses in avian louse flies from Germany to assess whether they could be used as sentinel organisms for monitoring arboviruses with zoonotic potential.
Methods:
We collected 1000 louse flies of the species Crataerina hirundinis, C. pallida, Ornithomya avicularia, O. biloba, O. fringillina, O. chloropus, Ornithophila metallica and Pseudolynchia canariensis in Germany and screened the samples via RT-PCR for West Nile virus (WNV), Usutu virus (USUV) and Sindbis virus (SINV), which are arboviruses with avian hosts as reservoirs.
Results:
While WNV was not detected, we found one louse fly positive for USUV and one for SINV RNA, both of which belonged to the species O. avicularia (n = 279). Therefore, the detection rates for both USUV and SINV were 0.1% (95% CI 0.0–0.3%) in the total sample and 0.36% (95% CI 0.00–1.09%) in O. avicularia. For the sample that tested positive for SINV, the PCR results were confirmed by sequencing a 288-bp segment that encoded part of the virus’s structural polyprotein.
Conclusions:
This is the first time that USUV RNA and SINV RNA have been detected in louse flies. In addition, it is the first detection of human pathogenic viruses in the louse fly species O. avicularia. The results of this study indicate that louse flies should not be neglected as possible sentinels of viral pathogens with zoonotic potential in the sense of the One Health concept
MLUH-E-116_1, Falco naumanni Fleischer, JG, 1818, eggshell
No full textInventory No.: MLUH-E-116_1, Object: eggshell, Species: Falco naumanni Fleischer, JG, 1818, Preservation: complete preservation, Locality_loc.: Smyrn Parnass, Locality today: Izmir/Parnassus, Country: Greece/TurkeyCollector_leg.: Dr. T. J. Krüper, Collection_coll.: M. Schönwetter, published in Handbuch der Oologie, Schönwetter, Max: Vol. I, p. 192, Identification by: M. Schönwetter, Aquisition: 1941, Aquired from: Dr. Henrici, ex Collection: ex. Coll. Dr. Henrici; ex. Coll. Dr. T. J. Krüper, Additional Information: 0.85
Essays on Option Pricing
No full textDerivative contracts play a fundamental role in the financial system since they provide unique flexibility and
precision in financial strategies, which are not possible with other instruments. Among these contracts, options
are the most traded derivative contracts worldwide, with a volume of 108.2 billion contracts in 2023 and, in the
first four months of 2024, their trading volume increased by 104 percent.1 The use of options has driven major
innovations in financial markets, enabling the development of instruments such as contingent claims, structured
and volatility derivatives. By replicating the pay-off of other financial assets, options contribute to market
efficiency through synthetic positions that take advantage of price inefficiencies across different assets and markets.
Furthermore, synthetic positions enhance market liquidity. For investors, options are essential tools in risk and
portfolio management, allowing them to hedge against asset volatility and leverage their positions. Option pricing
facilitates future estimates of the underlying asset volatility and reflects market expectations. The concept of
implied volatility2 which captures market uncertainty, serves as a robust measure for future volatility forecasting.
To sum up,options are a critical component of modern finance, and their role in the financial markets is crucial
for maintaining stability. The Black and Scholes (1973) model revolutionized options pricing by providing a
closed-form solution to value options. Since then, extensive research has built on the theoretical framework
of option pricing, price and return dynamics, empirical analysis, and forecasting. This cumulative dissertation
comprises four papers that contribute to this ongoing research by applying option pricing to company valuation,
exploring options rates of return dynamics from multiple perspectives, and analyzing options in the rapidly
evolving decentralized finance ecosystem.
The Merton (1973) model significantly transformed company valuation by representing equity and debt as options
on company’s assets. The first paper extends the computaion of the cost of capital within the Merton (1973) and
Modigliani and Miller (1958) frameworks incorporating credit risk. This framework also enables the calculation of
the value of the tax shield under credit risk. The paper critically discusses the debt beta approach where, under the
Capital Asset Pricing Model, the debt beta is calculated with respect to a combined market portfolio of stocks and
corporate bonds, which does not exist. Therefore, the first paper proposes an option-based and combined option-
factor sensitivities approach that integrates credit risk into the Weighted Average Cost of Capital computations.
The paper also compares the data requirements of the discussed approach and conducts a peer-group valuation
of Apple Inc. assuming it were not a publicly traded company.
Since the major research on options has been conducted for the US market, in the second and third papers, this
dissertation empirically analyzes options rates of return dynamics and compares the results on the US and the
EU markets. The second paper builds on Aretz et al. (2023) study and explores the relationship between equity
option rates of return and underlying volatility. The empirical study employs the Fama-French-Carhart and expo-
nential GARCH models to decompose the underlying volatility into systematic and idiosyncratic components and
examines their impact on option rates of return across different moneyness levels, including nonlinear relationship.
The paper provides detail discussion of systematic and idiosyncratic volatility slopes conditional on moneyness.
The results contribute to the literature by disclosing the nuanced impact of underlying systematic volatility on
option rates of return. The third paper extends the second by investigating the cross-section of index option
rates of return again on both the US and EU markets. It employs the ARIMA-GJR-GARCH model to estimate
the underlying index volatility and deploys two mixed effect panel regressions. The regression analysis focuses,
first, on the cross-section of option rates of return with respect to underlying volatility, volatility of volatility,
leverage, moneyness, and elasticity. Second, the analysis compares the elasticity sensitivities to its components,
i.e., delta and leverage, and discusses the elasticity dynamics across options and markets with respect to volatility
and moneyness bins. These papers contribute to the literature by providing empirical evidence on options market
dynamics and highlighting differences that should be taken into consideration for region-specific risk management.
Concluding this dissertation and with regard to a relatively new but rapidly growing financial market, the fourth
paper explores the pricing of cryptocurrency options on-chain as a part of a decentralized finance ecosystem. This
is the first empirical study in its field and investigates the pricing of wrapped Bitcoin and Ethereum options in the
Hegic protocol. It utilizes the two regimes Markov Switching Autoregression (GJR) GARCH model to estimate
the underlying volatility and the feasible GLS regression to examine the discrepancies between the Hegic and the
benchmark prices. The regression analysis includes the rate of return, volume, and volatility of the underlying
and options’ amount, strike, and moneyness. The paper also discusses the difference between implied and market
volatility, indicating potential mispricing exploitation. This paper contributes to volatility forecasting and opens
a discussion on a sophisticated automated market maker for pricing options in decentralized finance ecosystem.
Overall, this dissertation shows a manifold contribution to scientific research on option pricing. It covers both the
theoretical and empirical perspectives of option pricing, and the traditional and decentralized financial markets.
The results are practically relevant for risk assessment and management, investment strategies optimization, and
financial modeling.Literaturangabe
MLUH-E-150_1, Circus pygargus (Linnaeus, 1758), eggshell
No full textInventory No.: MLUH-E-150_1, Object: eggshell, Species: Circus pygargus (Linnaeus, 1758), Preservation: slight damage, Locality_loc.: Südrussland, Locality today: Southern Federal District, Country: RussiaCollection_coll.: M. Schönwetter, published in Handbuch der Oologie, Schönwetter, Max: Vol. I, p. 176, Identification by: M. Schönwetter, Aquisition: 1910, Aquired from: Dr. Rey, ex Collection: ex. Coll. Dr. Re
MLUH-E-283_2, Clanga pomarina (Brehm, CL, 1831), eggshell
No full textInventory No.: MLUH-E-283_2, Object: eggshell, Species: Clanga pomarina (Brehm, CL, 1831), Preservation: complete preservation, Locality_loc.: Birkhorst Mark Brandenburg, Locality today: Birkhorst, Country: GermanyCollector_leg.: W. Behrends, Date: 11/05/1891, Collection_coll.: M. Schönwetter, published in Handbuch der Oologie, Schönwetter, Max: Vol. I, p. 172, Identification by: M. Schönwetter, Aquisition: 1940, Aquired from: W. Behrends, ex Collection: ex. Coll. W. Behrend
MLUH-E-98_1, Falco tinnunculus Linnaeus, 1758, eggshell
No full textInventory No.: MLUH-E-98_1, Object: eggshell, Species: Falco tinnunculus Linnaeus, 1758, Preservation: complete preservation, Locality_loc.: Halle an der Saale, Locality today: Halle(Saale), Country: GermanyCollection_coll.: M. Schönwetter, published in Handbuch der Oologie, Schönwetter, Max: Vol. I, p. 192, Identification by: M. Schönwette