Ostbayerische Technische Hochschule
Recent publications
Background and aims: Celiac disease with its endoscopic manifestation of villous atrophy is underdiagnosed worldwide. The application of artificial intelligence (AI) for the macroscopic detection of villous atrophy at routine esophagogastroduodenoscopy may improve diagnostic performance. Methods: A dataset of 858 endoscopic images of 182 patients with villous atrophy and 846 images from 323 patients with normal duodenal mucosa was collected and used to train a ResNet 18 deep learning model to detect villous atrophy. An external data set was used to test the algorithm, in addition to six fellows and four board certified gastroenterologists. Fellows could consult the AI algorithm's result during the test. From their consultation distribution, a stratification of test images into "easy" and "difficult" was performed and used for classified performance measurement. Results: External validation of the AI algorithm yielded values of 90 %, 76 %, and 84 % for sensitivity, specificity, and accuracy, respectively. Fellows scored values of 63 %, 72 % and 67 %, while the corresponding values in experts were 72 %, 69 % and 71 %, respectively. AI consultation significantly improved all trainee performance statistics. While fellows and experts showed significantly lower performance for "difficult" images, the performance of the AI algorithm was stable. Conclusion: In this study, an AI algorithm outperformed endoscopy fellows and experts in the detection of villous atrophy on endoscopic still images. AI decision support significantly improved the performance of non-expert endoscopists. The stable performance on "difficult" images suggests a further positive add-on effect in challenging cases.
Artificial intelligence is steadily being integrated into all areas of medicine. In reproductive medicine, artificial intelligence methods can be utilized to improve the selection and prediction of sperm cells, oocytes, and embryos and to generate better predictive models for in vitro fertilization. The use of artificial intelligence in this field is justified by the suffering of persons or couples who wish to have children but are unable to conceive. However, research into the use of artificial intelligence in reproductive medicine is still in the early experimental stage and furthermore raises complex normative questions. There are ethical research challenges because evidence of the efficacy of certain pertinent systems is often lacking and because of the increased difficulty of ensuring informed consent on the part of the affected persons. Other ethically relevant issues include the potential risks for offspring and the difficulty of providing sufficient information. The opportunity to fulfill the desire to have children affects the welfare of patients and their reproductive autonomy. Ultimately, ensuring more accurate predictions and allowing physicians to devote more time to their patients will have a positive effect. Nevertheless, clinicians must be able to process patient data conscientiously. When using artificial intelligence, numerous actors are involved in making the diagnosis and deciding on the appropriate therapy, raising questions about who is ultimately responsible when mistakes occur. Questions of fairness arise with regard to resource allocation and cost reimbursement. Thus, before implementing artificial intelligence in clinical practice, it is necessary to critically examine the quantity and quality of the data used and to address issues of transparency. In the medium and long term, it would be necessary to confront the undesirable impact and social dynamics that may accompany the use of artificial intelligence in reproductive medicine.
Lima is the second-largest capital of the world located in a desert and already faces water scarcity. Here, more than 30% of the population is supplied by only 2.2% of the national water resources. The urbanization process has an informal nature and occurs at a very accelerated rate. These new settlements lack water infrastructure and access to other services. The objectives of this study are to quantify changes in the urban landscape of Lima Metropolitan Area from 1990 to 2021 to propose a water-sensitive urban plan by detecting changes, urbanization trends and identifying alternative water sources. The trend suggests a future constant increment of the urban areas, diversification of the landscape and more equally distributed land cover. Lima has more disconnected settlements and more complex shapes of urban patches nowadays. The landscape is also more mingled, but cracked. Overall, the trend is to become more disaggregated, demanding small and scattered water solutions. The WSUP includes the implementation of treatment plants in new multi-family buildings, hybrid desalination plants at the coast and parks with fog collectors on the hills. Additionally, these solutions will require the beneficiary community and the local authorities to work together in the planning and maintenance.
Background: Reconstructive surgery is established as a standard treatment option after mastectomy due to cancer. It is crucial to patients to achieve a natural and symmetric looking breast through reconstruction. Anthropometric measurements are used to assess the objective symmetry of the breast, which are prone to errors and difficult to reproduce. Objective: The aim of this work is to validate breast volumetry using three-dimensional surface imaging. Methods: We compared preoperatively analog and digitally evaluated volume of the breast with our gold standard, direct water displacement measurement of the mastectomy specimen. We examined 34 breast specimens in total. Results: Each measurement method (Breast Sculptor, VAM, Breast-V) for breast volume/mass determination demonstrates acceptable agreement ranges when compared with resected volumes and masses. The strongest volumetry instrument is Breast Sculptor (digital), the weakest is Breast-V (analog). Conclusions: 3D surface imaging is a quick, effective, and convenient method to evaluate breast shape and volume. The accuracy, reproducibility, and reliability of 3D surface imaging were comparable with MRI in our study.This takes us a step closer to the long-term goal of establishing robust instruments to plan breast reconstructive surgery, achieve better surgical results, and contribute to quality assurance in breast surgery.
Exoskeletons were invented over 100 years ago but have only become popular in the last two decades, especially in the working industry as they can decrease work-related loads significantly. The most often used exoskeletons are for the lower back and shoulder since these are commonly affected body regions. All devices have in common that their purpose is to reduce internal loads of vulnerable body regions. Nevertheless, there is still little understanding on how biomechanical loading in the human body changes when exoskeletons are used. Therefore, further analyses are needed. A promising candidate for these are musculoskeletal models, which are based on an inverse dynamics approach and can calculate external parameters such as ground reaction forces or other interaction forces as well as internal parameters such as joint reaction forces or muscle activities. The various examples in the literature show that these models are increasingly used for assessing the biomechanical effects of exoskeletons on the human body. Furthermore, musculoskeletal models can calculate biomechanical loadings of humans with and without exoskeletons for all kinds of applications and allow an evaluation of their purpose. Practical Relevance: This article highlights the possibilities of musculoskeletal models for assessing the design and efficiency of occupational exoskeletons. Several practical use cases are described along with distinct descriptions of common implications of musculoskeletal and exoskeleton modeling.
This paper introduces a novel chaotic flower pollination algorithm (CFPA) to solve a tardiness-constrained flow-shop scheduling problem with simultaneously loaded stations. This industrial manufacturing problem is modeled from a filter basket production line in Germany and has been generally solved using standard deterministic algorithms. This research develops a metaheuristic approach based on the highly efficient flower pollination algorithm coupled with different chaos maps for stochasticity. The objective function targeted is the tardiness constraint of the due dates. Fifteen different experiments with thirty scenarios are generated to mimic industrial conditions. The results are compared with the genetic algorithm and with the four standard benchmark priority rule-based deterministic algorithms of First In First Out, Raghu and Rajendran, Shortest Processing Time and Slack. From the obtained results and analysis of the relative difference, percentage relative difference and t tests, CFPA was found to be significantly better performing than the deterministic heuristics and the GA algorithm.
We present a comprehensive analysis of the neural audio-visual synchrony evaluation tool SyncNet. We assess the agreement of SyncNet scores vis-a-vis human perception and whether we can use these as a reliable metric for evaluating audio-visual lip-synchrony in generation tasks with no ground truth reference audio-video pair. We further look into the underlying elements in audio and video which vitally affect synchrony using interpretable explanations from SyncNet predictions and analyse its susceptibility by introducing adversarial noise. SyncNet has been used in numerous papers on visually-grounded text-to-speech for scenarios such as dubbing. We focus on this scenario which features many local asynchronies (something that SyncNet isn’t made for).
Zusammenfassung In der Vergangenheit wurde beklagt, dass in der empirischen Forschung zur Bildungsbeteiligung Erwachsener selten eine Integration von Theorie und Empirie erfolge und die mangelnde empirische Prüfung eine theoretische Weiterentwicklung des Forschungsfeldes erschwere. Mittels einer Qualitativen Inhaltsanalyse von Beiträgen ausgewählter Fachzeitschriften der Erwachsenen‑/Weiterbildungsforschung wurde versucht, die Aktualität dieser Feststellung empirisch zu überprüfen. Dazu wurden 38 empirische Forschungsbeiträge zur Bildungsbeteiligung der Jahrgänge 2018 bis 2020 von neun Fachzeitschriften mit Peer-Review untersucht. Die Einschätzung der Beiträge erfolgte hinsichtlich des Grades der Theorieeinbindung. Die Ergebnisse der Analyse zeigen ein gemischtes Bild: neun Beiträge (24 %) binden keine Theorie in den Forschungsprozess ein, ein Beitrag (3 %) zitiert Theorie lediglich, ohne sie weiter einzubinden. Am häufigsten ist eine Einbindung zum theoriegeleiteten Vorgehen (16 Beiträge, 42 %). Eine tiefergehende Theorieanwendung findet sich in sechs Beiträgen (16 %). Eine theorieprüfende oder -bildende Einbindung liegt mit je drei Beiträgen (je 8 %) seltener vor. Die rezipierten Theorien sind vielfältig: es werden 20 verschiedene Theorieansätze festgestellt. Den Forderungen nach einer verstärkten Einbindung von Theorie in den Forschungsprozess kann somit weiterhin Gültigkeit attestiert werden: ein substanzieller Teil der Beiträge berichtet keine Theorieeinbindung, eine empirische Prüfung oder Weiterentwicklung findet relativ selten statt.
Modern cyber-physical systems, such as autonomous vehicles, advanced driver assistance systems, automation systems and battery management systems, result in extended communication requirements regarding the reliability and the availability. The Controller Area Network (CAN) is a broadcast-based protocol which is still used as a standard for serial communication between individual microcontrollers due to its reliability and low power consumption. In addition, it provides mechanisms for detecting transmission errors and retransmitting messages in the event of an error. The enhancement CAN Flexible Data-Rate (CAN FD) offers increased data rates and transmission rates in order to meet the data throughput requirements. In this paper, the mechanisms for reliable data transmission in a CAN FD network are analyzed. To improve reliability, a second identical CAN-FD network is added to the system, using the additional CAN interface already available on common microcontrollers. The redundant communication network is examined in terms of failure rates and the mean time to failure. The reliability over the operation time is calculated for the single and the redundant version of the CAN FD network using the failure rate limits of the ASIL levels.
The laminated steel core of transformers is one of the main sources of the generated sound, as it is excited by different electromagnetic effects during its normal operation. If the core is excited in its eigenfrequencies, the sound generated by a transformer will increase significantly. Therefore, knowledge of the core’s eigenmodes and -frequencies in an early design stage can decrease expenses by avoiding costly modifications that might be required to avoid the sound levels exceeding the specified values after the final factory acceptance test. The current study focuses on developing a core resonance model to determine the eigenmodes and -frequencies of a transformer core. The core’s geometry was simplified to a connected-beam structure and a numerical-based approach was applied. The accuracy of the developed model was validated against finite element method (FEM), using ANSYS on a reference core model.
During the factory acceptance test, the sound levels are measured. If the measured sound levels exceed pre-specified values, modifications on the transformer’s active part and/or tank need to be conducted. Tank resonance, if occurs, at twice main power frequency and its higher harmonics is one of the main reasons of amplifying the generated sound levels. Therefore, it is preferable to know already in the design stage about the occurrence of the tank resonance and propose reliable tank solutions. The current study presents a newly developed model for calculating the mode shapes and its corresponding eigenfrequencies of the tank.
Background: Classic (static) Roentgen stereophotogrammetric analysis (RSA) is the current gold standard to assess, in vivo, the migration of total joint arthroplasty components. To prevent potential patient motion artifacts during the acquisition of paired radiostereometric images, images must be taken by simultaneously firing both X-ray tubes. However, the influence of nonsynchronized RSA paired images or patient motion artifacts on the precision of RSA and the assessment of implant migration is not well understood. Questions/purposes: We assessed (1) the effect of possible patient motion on the precision of RSA and (2) apparent differences in implant migration among axes (in-plane and out-of-plane translations and in-plane and out-of-plane rotations) of possible motion artifacts. Methods: Radiographs of two tibial knee arthroplasty components, each fixed in two bone-implant models as a customized phantom, were taken in a uniplanar measurement setup. We evaluated both model-based (implant models from reversed engineering) and marker-based (additional attached implant markers) RSA approaches. Between the simulated reference and follow-up examinations, we used one of the bone-implant models to simulate patient motion and the other to simulate no patient motion in parallel. Two defined protocols were followed for each of the bone-implant models: no-motion and simulated motion protocols. RSA image pairs were analyzed using a model-based RSA software package (MBRSA 4.1, RSAcore). Precision was calculated through repeat examinations, and migration of the two components was assessed for comparison of the components with each other. Measurements were taken along the medial-lateral and posterior-anterior axes for translations and around the cranial-caudal axis for rotations. The maximum total point motion was measured for comparison between the two components. Results: The effect of simulated patient motion was generally small, except in the cranial-caudal axis, but the induced imprecision associated with motion was larger in model-based RSA than it was in marker-based RSA. The mean ± standard deviation values of precision in model-based RSA were 0.035 ± 0.015 mm, 0.045 ± 0.014 mm, and 0.049 ± 0.036 mm greater than those in marker-based RSA, in accordance with the simulated motion protocol in translations along the medial-lateral axis (0.018 ± 0.004 mm; p = 0.01), along the posterior-anterior axis (0.018 ± 0.007 mm; p = 0.003), and rotations around the cranial-caudal axis (0.017 ± 0.006 mm; p = 0.02). Apparent differences in implant migration were the greatest for the maximum total point motion. The maximum total point motion increased from 0.038 ± 0.007 mm for the no-motion protocol to 1.684 ± 0.038 mm (p < 0.001) for the simulated motion protocol in marker-based RSA, and from 0.101 ± 0.027 mm for the no-motion protocol to 1.973 ± 0.442 mm (p < 0.001) for the simulated motion protocol in model-based RSA, and was the worst-case scenario regarding patient motion artifacts. Conclusion: Patient motion exceeding 1 mm or 1° on nonsynchronized RSA images affects measurement errors regarding the detection of migration of a tibial component. In clinical RSA studies, the effect of patient motion on the assessment of implant migration should be of particular concern, even if clinical RSA systems have acceptable precision. Specially trained radiographers are crucial for correctly acquiring radiographs, especially when simultaneous radiography exposures are not electronically automated. In general, RSA requires synchronized image acquisition, and this should be the state-of-the-art. Clinical relevance: In clinical RSA studies, precision assessed by repeat examinations may not be reliable using the current standards that are widely used in radiology departments. When assessing implant migration for reliability, comparison of the maximum total point motion between the tested (simulated motion) implant and baseline (no-motion) implant, as in this study, is advocated because of the accurate detection of patient motion artifacts.
We proof a Garrett–Böcherer decomposition of a vector-valued Siegel Eisenstein series El,02\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_{l,0}^2$$\end{document} of genus 2 transforming with the Weil representation of Sp2(Z)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\text {Sp}}_2({\mathbb {Z}})$$\end{document} on the group ring C[(L′/L)2]\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathbb {C}}[(L'/L)^2]$$\end{document}. We show that the standard zeta function associated to a vector-valued common eigenform f for the Weil representation can be meromorphically continued to the whole s-plane and that it satisfies a functional equation. The proof is based on an integral representation of this zeta function in terms of f and El,02\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_{l,0}^2$$\end{document}.
Jedes Jahr erleiden 270.000 Menschen in Deutschland einen Schlaganfall. In vielen Fällen können die betroffenen Personen wieder nach Hause zurückkehren und ihr Leben fortführen, benötigen dabei aber ambulante Pflege- und Therapiemaßnahmen. Gerade in ländlichen Regionen bringt dies erhebliche Herausforderungen mit sich, denen sich neue Technologien und die Digitalisierung entgegenstellen. Die Beiträger*innen des Bandes diskutieren erste Ergebnisse des Projekts »DeinHaus 4.0 Oberpfalz« aus interdisziplinärer Sicht, bei dem die Möglichkeit des Einsatzes von Telepräsenzrobotern zur Unterstützung ambulanter Pflege- und Therapiemaßnahmen untersucht wird.
Jedes Jahr erleiden 270.000 Menschen in Deutschland einen Schlaganfall. In vielen Fällen können die betroffenen Personen wieder nach Hause zurückkehren und ihr Leben fortführen, benötigen dabei aber ambulante Pflege- und Therapiemaßnahmen. Gerade in ländlichen Regionen bringt dies erhebliche Herausforderungen mit sich, denen sich neue Technologien und die Digitalisierung entgegenstellen. Die Beiträger*innen des Bandes diskutieren erste Ergebnisse des Projekts »DeinHaus 4.0 Oberpfalz« aus interdisziplinärer Sicht, bei dem die Möglichkeit des Einsatzes von Telepräsenzrobotern zur Unterstützung ambulanter Pflege- und Therapiemaßnahmen untersucht wird.
Jedes Jahr erleiden 270.000 Menschen in Deutschland einen Schlaganfall. In vielen Fällen können die betroffenen Personen wieder nach Hause zurückkehren und ihr Leben fortführen, benötigen dabei aber ambulante Pflege- und Therapiemaßnahmen. Gerade in ländlichen Regionen bringt dies erhebliche Herausforderungen mit sich, denen sich neue Technologien und die Digitalisierung entgegenstellen. Die Beiträger*innen des Bandes diskutieren erste Ergebnisse des Projekts »DeinHaus 4.0 Oberpfalz« aus interdisziplinärer Sicht, bei dem die Möglichkeit des Einsatzes von Telepräsenzrobotern zur Unterstützung ambulanter Pflege- und Therapiemaßnahmen untersucht wird.
Institution pages aggregate content on ResearchGate related to an institution. The members listed on this page have self-identified as being affiliated with this institution. Publications listed on this page were identified by our algorithms as relating to this institution. This page was not created or approved by the institution. If you represent an institution and have questions about these pages or wish to report inaccurate content, you can contact us here.
2,861 members
Christoph Palm
  • Computer Science
Markus Westner
  • Computer Science and Mathematics
Charlotte Thiel
  • Civil Engineering
Jan Dünnweber
  • Computer Science and Mathemetics
Markus Kucera
  • Computer Science & Mathematics
Pruefeninger Str. 58, 93049, Regensburg, Germany
Head of institution
Prof. Dr. Wolfgang Baier