05 Fakultät Informatik, Elektrotechnik und Informationstechnik
Permanent URI for this collectionhttps://elib.uni-stuttgart.de/handle/11682/6
Browse
12 results
Search Results
Item Open Access A muscle model for injury simulation(2023) Millard, Matthew; Kempter, Fabian; Fehr, Jörg; Stutzig, Norman; Siebert, TobiasCar accidents frequently cause neck injuries that are painful, expensive, and difficult to simulate. The movements that lead to neck injury include phases in which the neck muscles are actively lengthened. Actively lengthened muscle can develop large forces that greatly exceed the maximum isometric force. Although Hill-type models are often used to simulate human movement, this model has no mechanism to develop large tensions during active lengthening. When used to simulate neck injury, a Hill model will underestimate the risk of injury to the muscles but may overestimate the risk of injury to the structures that the muscles protect. We have developed a musculotendon model that includes the viscoelasticity of attached crossbridges and has an active titin element. In this work we evaluate the proposed model to a Hill model by simulating the experiments of Leonard et al. [1] that feature extreme active lengthening.Item Open Access Editorial - autonomous health monitoring and assistance systems with IoT(2021) Azzopardi, George; Karastoyanova, Dimka; Aiello, Marco; Schizas, Christos N.Item Open Access Improving the accuracy of musculotendon models for the simulation of active lengthening(2023) Millard, Matthew; Kempter, Fabian; Stutzig, Norman; Siebert, Tobias; Fehr, JörgVehicle accidents can cause neck injuries which are costly for individuals and society. Safety systems could be designed to reduce the risk of neck injury if it were possible to accurately simulate the tissue-level injuries that later lead to chronic pain. During a crash, reflexes cause the muscles of the neck to be actively lengthened. Although the muscles of the neck are often only mildly injured, the forces developed by the neck’s musculature affect the tissues that are more severely injured. In this work, we compare the forces developed by MAT_156, LS-DYNA’s Hill-type model, and the newly proposed VEXAT muscle model during active lengthening. The results show that Hill-type muscle models underestimate forces developed during active lengthening, while the VEXAT model can more faithfully reproduce experimental measurements.Item Open Access ‘Better see a doctor?’ status quo of symptom checker apps in Germany : a cross-sectional survey with a mixed-methods design (CHECK.APP)(2024) Wetzel, Anna-Jasmin; Koch, Roland; Koch, Nadine; Klemmt, Malte; Müller, Regina; Preiser, Christine; Rieger, Monika; Rösel, Inka; Ranisch, Robert; Ehni, Hans-Jörg; Joos, StefanieBackground: Symptom checker apps (SCAs) offer symptom classification and low-threshold self-triage for laypeople. They are already in use despite their poor accuracy and concerns that they may negatively affect primary care. This study assesses the extent to which SCAs are used by medical laypeople in Germany and which software is most popular. We examined associations between satisfaction with the general practitioner (GP) and SCA use as well as the number of GP visits and SCA use. Furthermore, we assessed the reasons for intentional non-use. Methods: We conducted a survey comprising standardised and open-ended questions. Quantitative data were weighted, and open-ended responses were examined using thematic analysis. Results: This study included 850 participants. The SCA usage rate was 8%, and approximately 50% of SCA non-users were uninterested in trying SCAs. The most commonly used SCAs were NetDoktor and Ada. Surprisingly, SCAs were most frequently used in the age group of 51–55 years. No significant associations were found between SCA usage and satisfaction with the GP or the number of GP visits and SCA usage. Thematic analysis revealed skepticism regarding the results and recommendations of SCAs and discrepancies between users’ requirements and the features of apps. Conclusion: SCAs are still widely unknown in the German population and have been sparsely used so far. Many participants were not interested in trying SCAs, and we found no positive or negative associations of SCAs and primary care.Item Open Access Deep learning aided clinical decision support(2023) Schneider, Rudolf; Staab, Steffen (Prof. Dr.)Medical professionals create vast amounts of clinical texts during patient care. Often, these documents describe medical cases from anamnesis to the final clinical outcome. Automated understanding and selection of relevant medical records pose an opportunity to assist medical doctors in their day-to-day work on a large scale. However, clinical text understanding is challenging, especially when dealing with clinical narratives such as nursing notes or diagnostic reports. These clinical documents differ extensively in length, structure, vocabulary, and lexical and grammatical correctness. In addition, they are highly context-dependent. For all these reasons, approaches based on syntactic rules and discrete text representation often fail to address the variety of clinical narratives propagating unrecoverable errors to downstream applications. Therefore, this thesis focuses on evaluating and designing methods and models that are generalizable and adaptable enough to deal with these challenges. Our goal is to enable text-based clinical decision support systems to utilize the knowledge from clinical archives and medical publications. We aim to design methods that can scale up to the growing amount of clinical documents in hospital archives. A fundamental problem in achieving deep-learning-enabled clinical decision support systems is designing a patient representation that captures all relevant information for automated processing. We engage these challenges by designing a framework for deep-learning-enabled differential diagnosis support. Guided by the needs emerging from this framework, we design and evaluate methods based on three information representation paradigms: (1) Discrete relation extraction using the open information extraction paradigm. (2) Neural text representations based on language and topic modeling. (3) Combining complementary neural text representations. Our framework translates clinical diagnostic steps and pathways to statistical and deep-learning-based models. Accordingly, we can show that deep-learning-enabled differential diagnosis benefits from contextualized information representations. Further, we identify shortcomings of the open information extraction paradigm in a comprehensive benchmark. We design a distributed text representation model based on topical information. Our extensive large-scale experiment results show that topical distributed text representations capture information complementary to language modeling-based approaches across domains, thus enabling a holistic text representation for medical texts. Our experiments with medical doctors using our prototypical implementation of the deep-learning-enabled differential diagnosis process validate this framework. Moreover, we identify seven crucial design challenges for text-based clinical decision support systems based on our qualitative and quantitative findings.Item Open Access Cervical muscle reflexes during lateral accelerations(2023) Millard, Matthew; Hunger, Susanne; Broß, Lisa; Fehr, Jörg; Holzapfel, Christian; Stutzig, Norman; Siebert, TobiasAutonomous vehicles will allow a variety of seating orientations that may change the risk of neck injury during an accident. Having a rotated head at the time of a rear-end collision in a conventional vehicle is associated with a higher risk of acute and chronic whiplash. The change in posture affects both the movement of the head and the response of the muscles. We are studying the reflexes of the muscles of the neck so that we can validate the responses of digital human body models that are used in crash simulations. The neck movements and muscle activity of 21 participants (11 female) were recorded at the Stuttgart FKFS mechanical driving simulator. During the maneuver we recorded the acceleration of the seat and electromyographic (EMG) signals from the sternocleidomastoid (STR) muscles using a Biopac MP 160 system (USA). As intuition would suggest, the reflexes of the muscles of the neck are sensitive to posture and the direction of the acceleration.Item Open Access Whiplash simulation: how muscle modelling and movement interact(2022) Millard, Matthew; Siebert, Tobias; Stutzig, Norman; Fehr, JörgWhiplash injury and associated disorders are costly to society and individuals. Accurate simulations of neck movement during car accidents are needed to assess the risk of whiplash injury. Existing simulations indicate that Hill-type muscle models are too compliant, and as a result, predict more neck movement than is observed during in-vivo experiments. Simulating head and neck movement is challenging because many of the neck muscles operate on the descending limb of the force-length curve, a region that Hill-type models inaccurately capture. Hill-type muscle models have negative stiffness on the descending limb of the force-length curve and so develop less force the more they are lengthened. Biological muscle, in contrast, can develop large transient forces during active lengthening and sustain large forces when aggressively lengthened. Recently, a muscle model has been developed that mimics the active impedance of muscle in the short range and can capture the large forces generated during extreme lengthening. In this work, we will compare the accuracy of simulated neck movements, using both a Hill-type model and the model of Millard et al., to the in-vivo neck movement. If successful, the improved accuracy of our simulations will make it possible to predict and help prevent neck injury.Item Open Access Avoiding shortcut-learning by mutual information minimization in deep learning-based image processing(2023) Fay, Louisa; Cobos, Erick; Yang, Bin; Gatidis, Sergios; Küstner, ThomasItem Open Access Optische Messsysteme und Ein-Sensor-Bildgebungsverfahren für Biosensoren(2024) Berner, Marcel; Werner, Jürgen H. (Prof. Dr. rer. nat. habil.)Die vorliegende Arbeit präsentiert die Entwicklung mehrerer Messsysteme und -verfahren für optische Biosensoranwendungen. Der erste Teil dieser Arbeit entwirft eine universelle experimentelle Plattform für die Erprobung neuer optischer Biosensorkonzepte nach dem Prinzip der laserinduzierten Fluoreszenz (LIF). Die Plattform unterstützt das europäische Forschungsprojekt Nanodem bei der Entwicklung eines portablen Point-of-Care-Testing-Gerätes (PoCT) zur Live-Überwachung von Immunsuppressivakonzentrationen im Blut von Transplantationspatienten unmittelbar am Patientenbett. Das in dieser Arbeit entwickelte Plattformkonzept umfasst die optoelektronische Fluoreszenzanregung und -detektion, optische Filtersysteme, den fluoreszenten Farbstoff, das Materialsystem der Transducerchips, das Mikrofluidiksystem sowie die Automatisierung der Ablaufsteuerung. Der Ausgangspunkt der Entwicklung ist die Herleitung eines allgemeinen physikalischen Modells für LIF-Systeme, an dem sich die Konstruktion der Plattform orientiert. Das in Kooperation mit der Eberhard Karls Universität Tübingen entworfene Transducerchipkonzept auf der Basis lasergeschnittener Klebebänder gestattet eine hohe Flexibilität bezüglich der Geometrie und des Aufbaus der Transducerchips und unterstützt den Technologietransfer akademischer Forschungsergebnisse in die industrielle Fertigung. Die entworfenen Photodetektorarrays aus amorphem Silizium lassen sich dank leicht adaptierbarer Herstellungsprozesse kosteneffizient auf beliebige Biosensorgeometrien anpassen. Die erreichte spezifische Detektivität D* = 11 × 10^12 Jones der Detektoren liegt dabei auf Augenhöhe mit der von State-of-the-Art-Detektoren aus kristallinem Material. Die erzielte Detektionsgrenze von c_{LOD,exp} = 26 nmol/l. Weiter bestätigen die experimentellen Messdaten das aufgestellte physikalische Modell. Der zweite Teil dieser Arbeit zeigt ein neues optisches Verfahren zur ortsaufgelösten Messung, das eine Vielzahl von Bildpunkten simultan mit nur einem einzigen optischen Sensor beobachtet. Das Verfahren nutzt hierzu ortsaufgelöste Lichtmodulatoren (Spatial Light Modulators - SLMs), um eine ortsabhängige optische Modulation zu erzeugen. Die erzeugten optischen Trägersignale gestatten die Zuordnung der als Summensignal empfangenen Signale zu ihren Ursprungspunkten. Der sogenannte Fourier Spotter macht sich dabei die mathematischen Eigenschaften der Fourier-Transformation zunutze. Durch die Anwendung zueinander phasenverschobener Modulationssignale gestattet der Fourier Spotter zudem die unmittelbare Messung von Helligkeitsdifferenzen zwischen unterschiedlichen Beobachtungspunkten. Dieses differentielle optische Messprinzip ist der Kern eines bereits erteilten Patents des Autors mit der Universität Stuttgart. Das neuartige optische Messprinzip eignet sich für die Integration in optische Biosensor-Verfahren, wie etwa die Einwellenlängenreflektometrie (engl. Single Color Reflectometry - SCORE), welche derzeit noch auf teure Spezialkameras angewiesen sind. Herkömmliche Kamerasysteme erzeugen hohe Datenmengen, deren Auswertung erhebliche Rechenleistung in Anspruch nimmt und damit der Weiterentwicklung hin zu miniaturisierten, portablen Biosensorplattformen entgegensteht. Die vorliegende Arbeit präsentiert einen erfolgreichen experimentellen Machbarkeitsnachweis des Fourier Imagers anhand von Helligkeitsdifferenzmessungen an einem SCORE-Aufbau. Eine zukünftige Erweiterung des Fourier Spotters um ein Zeilenspektrometer erlaubt neben der ortsaufgelösten Beobachtung auch eine simultane Erfassung der optischen Spektren jedes einzelnen beobachteten Punktes. Durch diese hyperspektrale Erweiterung wird die erstmalige Umsetzung einer auf der reflektometrischen Interferenzspektroskopie (RIfS) basierenden mehrkanaligen optischen Biosensorplattform möglich. Der dritte Teil dieser Arbeit verallgemeinert das Prinzip des Fourier Spotters und überführt dieses in ein Ein-Pixel-Kamera-Verfahren - das AM-FDM Imaging (engl. Amplitude Modulated Frequency Division Multiplexing). Das AM-FDM Imaging basiert auf der Anwendung von Näherungsverfahren, die ein Übersprechen zwischen den Trägersignalen minimieren. Das aufgestellte systemtheoretische Modell des AM-FDM Imaging umfasst auch das Fourier Spotting und erlaubt den Vergleich mit Rasterscans sowie bereits bekannten Ein-Pixel-Kamera-Verfahren wie dem Hadamard Imaging. Ist das Signal-zu-Rausch-Verhältnis durch das Rauschen des Detektorsystems begrenzt, so erreicht das AM-FDM Imaging einen sogenannten Multiplexgewinn amult = O(M) in der Größenordnung der Anzahl simultan beobachteter Bildpunkte M. Mit den derzeit eingesetzten Näherungsverfahren erreicht das AM-FDM Imaging hinsichtlich des Signal-zu-Rausch-Verhältnisses, der Anzahl simultan beobachtbarer Bildpunkte und der erzielbaren Bildwiederholrate nicht die Leistungsfähigkeit des bei Ein-Pixel-Imaging-Verfahren vorherrschenden Hadamard Imagings. Die in dieser Arbeit diskutierten Verwandtschaftsverhältnisse des AM-FDM Imagings zu anderen bekannten Ein-Pixel-Kamera-Verfahren legen jedoch die Vermutung nahe, dass ein bisher unbekanntes Näherungsverfahren existiert, das das AM-FDM Imaging mit dem Hadamard Imaging gleichstellt. Die Ergebnisse des systemtheoretischen Modells wurden mittels Simulation in Matlab bestätigt und gelten auch für den Fourier Spotter. Damit zeigen die Ergebnisse auf, dass im SCORE-Anwendungsfall eine Modulation nach dem Prinzip des Hadamard Imagings vorteilhafter ist. Das erteilte Patent zum optisch differentiellen Messverfahren schließt auch eine differentielle Variante des Hadamard Imagings mit ein. Gegenüber der Differenzwertbestimmung aus gemessenen Absolutwerten verdoppelt das differentielle Messverfahren wahlweise das Signal-zu-Rauschleistungs-Verhältnis oder die Bildwiederholrate des Hadamard Imagings.Item Open Access Automated imaging-based abdominal organ segmentation and quality control in 20,000 participants of the UK Biobank and German National Cohort Studies(2022) Kart, Turkay; Fischer, Marc; Winzeck, Stefan; Glocker, Ben; Bai, Wenjia; Bülow, Robin; Emmel, Carina; Friedrich, Lena; Kauczor, Hans-Ulrich; Keil, Thomas; Kröncke, Thomas; Mayer, Philipp; Niendorf, Thoralf; Peters, Annette; Pischon, Tobias; Schaarschmidt, Benedikt M.; Schmidt, Börge; Schulze, Matthias B.; Umutle, Lale; Völzke, Henry; Küstner, Thomas; Bamberg, Fabian; Schölkopf, Bernhard; Rückert, Daniel; Gatidis, SergiosLarge epidemiological studies such as the UK Biobank (UKBB) or German National Cohort (NAKO) provide unprecedented health-related data of the general population aiming to better understand determinants of health and disease. As part of these studies, Magnetic Resonance Imaging (MRI) is performed in a subset of participants allowing for phenotypical and functional characterization of different organ systems. Due to the large amount of imaging data, automated image analysis is required, which can be performed using deep learning methods, e. g. for automated organ segmentation. In this paper we describe a computational pipeline for automated segmentation of abdominal organs on MRI data from 20,000 participants of UKBB and NAKO and provide results of the quality control process. We found that approx. 90% of data sets showed no relevant segmentation errors while relevant errors occurred in a varying proportion of data sets depending on the organ of interest. Image-derived features based on automated organ segmentations showed relevant deviations of varying degree in the presence of segmentation errors. These results show that large-scale, deep learning-based abdominal organ segmentation on MRI data is feasible with overall high accuracy, but visual quality control remains an important step ensuring the validity of down-stream analyses in large epidemiological imaging studies.