Hybrid Open Access - OA-Veröffentlichung in einer Subskriptionszeitschrift/-medium
Refine
Year of publication
Document Type
- Article (143)
- Part of a Book (1)
- conference proceeding (article) (1)
- conference talk (1)
- conference proceeding (volume) (1)
- Review (1)
Is part of the Bibliography
- no (148)
Keywords
- Shotcrete (4)
- 3D volumetry (3)
- Artificial Intelligence (3)
- Biomechanik (3)
- Durability (3)
- Maschinelles Lernen (3)
- Medical Image Computing (3)
- Photoacoustic spectroscopy (3)
- Acetone breath analysis (2)
- Acetone detection (2)
Institute
- Fakultät Informatik und Mathematik (47)
- Fakultät Maschinenbau (44)
- Regensburg Center of Biomedical Engineering - RCBE (22)
- Fakultät Elektro- und Informationstechnik (18)
- Regensburg Center of Health Sciences and Technology - RCHST (18)
- Fakultät Angewandte Sozial- und Gesundheitswissenschaften (16)
- Regensburg Medical Image Computing (ReMIC) (15)
- Fakultät Bauingenieurwesen (14)
- Fakultät Angewandte Natur- und Kulturwissenschaften (13)
- Labor Biomechanik (LBM) (12)
Begutachtungsstatus
- peer-reviewed (137)
Migration and Statistics
(2010)
Background
The purpose of this study was to evaluate the impact of Cone Beam CT (CBCT) based setup correction on total dose distributions in fractionated frameless stereotactic radiation therapy of intracranial lesions.
Methods
Ten patients with intracranial lesions treated with 30 Gy in 6 fractions were included in this study. Treatment planning was performed with Oncentra® for a SynergyS® (Elekta Ltd, Crawley, UK) linear accelerator with XVI® Cone Beam CT, and HexaPOD™ couch top. Patients were immobilized by thermoplastic masks (BrainLab, Reuther). After initial patient setup with respect to lasers, a CBCT study was acquired and registered to the planning CT (PL-CT) study. Patient positioning was corrected according to the correction values (translational, rotational) calculated by the XVI® system. Afterwards a second CBCT study was acquired and registered to the PL-CT to confirm the accuracy of the corrections. An in-house developed software was used for rigid transformation of the PL-CT to the CBCT geometry, and dose calculations for each fraction were performed on the transformed CT. The total dose distribution was achieved by back-transformation and summation of the dose distributions of each fraction. Dose distributions based on PL-CT, CBCT (laser set-up), and final CBCT were compared to assess the influence of setup inaccuracies.
Results
The mean displacement vector, calculated over all treatments, was reduced from (4.3 ± 1.3) mm for laser based setup to (0.5 ± 0.2) mm if CBCT corrections were applied. The mean rotational errors around the medial-lateral, superior-inferior, anterior-posterior axis were reduced from (−0.1 ± 1.4)°, (0.1 ± 1.2)° and (−0.2 ± 1.0)°, to (0.04 ± 0.4)°, (0.01 ± 0.4)° and (0.02 ± 0.3)°. As a consequence the mean deviation between planned and delivered dose in the planning target volume (PTV) could be reduced from 12.3% to 0.4% for D95 and from 5.9% to 0.1% for Dav. Maximum deviation was reduced from 31.8% to 0.8% for D95, and from 20.4% to 0.1% for Dav.
Conclusion
Real dose distributions differ substantially from planned dose distributions, if setup is performed according to lasers only. Thermoplasic masks combined with a daily CBCT enabled a sufficient accuracy in dose distribution.
We consider the reconstruction problem for limited angle tomography using filtered backprojection (FBP) and lambda tomography. We use microlocal analysis to explain why the well-known streak artifacts are present at the end of the limited angular range. We explain how to mitigate the streaks and prove that our modified FBP and lambda operators are standard pseudodifferential operators, and so they do not add artifacts. We provide reconstructions to illustrate our mathematical results.
Software engineering in open source projectsfaces similar challenges as in traditional software development(coordination of and cooperation between contributors, changeand release management, quality assurance, . . .), but often usesdifferent means of solving them. This leads to some salientdistinctions between both worlds, especially with respect tocommunication and how technical issues are addressed. Thevariations within open source software (OSS) communities areconsiderable, and many different approaches are currently inuse, ranging from informal conventions to highly systematic,formally specified and rigidly applied processes. We discussthe archetypal best practises in the field, illustrate them bypresenting example projects, and provide a comparison to tradi-tional approaches.
Background:
Over the past ten years, we carried out several comprehensive studies in the area of entrepreneurship education using business games that simulate the business processes of startups. The studies researched the learning effects of these games on German university students.
Aim:
The studies we present investigated if the simulation game-based educational entrepreneurship programs in Germany reached their goals. These goals included the fostering of participants’ entrepreneurial competencies and desire to create a startup.
Method:
We show traditional summative output-oriented results and more profound results based on our theory-based evaluation approach. A so-called logic model provides a framework for the interpretation of what takes place during the entrepreneurship business simulations. Currently, we have conducted research on more than 50 regular university courses with more than 1000 students. Another part of our studies is the evaluation of entrepreneurship education in the German national business game-based competition, EXIST priME Cup (further called EPC). More than 12,000 students attended more than 500 of these cup seminars. We also show new results from an online survey that was carried out to measure the long-term effects of the startup simulation-based seminars.
Results:
The results of five studies are summarized, and show an overall increase in the participants’ knowledge of business administration and business plan preparation skills (required in start-up context). Apart from other results, significant gender-based differences are identified. The differences relate to entrepreneurial attitudes and motivation. Some test scenarios with cup seminars for women only with extended debriefing seem to reduce this decrease in the motivation of women; further research with larger samples is necessary. A new study on the long-term effects shows a very high acceptance rate and demonstrates the satisfaction of former students. The simulation game-based program resulted in participants starting startups at a rate that is approximately twice as high (around 16%) as the normal rate in Germany.
Ductile failure of structural metals is relevant to a wide range of engineering scenarios. Computational methods are employed to anticipate the critical conditions of failure, yet they sometimes provide inaccurate and misleading predictions. Challenge scenarios, such as the one presented in the current work, provide an opportunity to assess the blind, quantitative predictive ability of simulation methods against a previously unseen failure problem. Rather than evaluate the predictions of a single simulation approach, the Sandia Fracture Challenge relies on numerous volunteer teams with expertise in computational mechanics to apply a broad range of computational methods, numerical algorithms, and constitutive models to the challenge. This exercise is intended to evaluate the state of health of technologies available for failure prediction. In the first Sandia Fracture Challenge, a wide range of issues were raised in ductile failure modeling, including a lack of consistency in failure models, the importance of shear calibration data, and difficulties in quantifying the uncertainty of prediction [see Boyce et al. (Int J Fract 186:5-68, 2014) for details of these observations]. This second Sandia Fracture Challenge investigated the ductile rupture of a Ti-6Al-4V sheet under both quasi-static and modest-rate dynamic loading (failure in 0.1 s). Like the previous challenge, the sheet had an unusual arrangement of notches and holes that added geometric complexity and fostered a competition between tensile- and shear-dominated failure modes. The teams were asked to predict the fracture path and quantitative far-field failure metrics such as the peak force and displacement to cause crack initiation. Fourteen teams contributed blind predictions, and the experimental outcomes were quantified in three independent test labs. Additional shortcomings were revealed in this second challenge such as inconsistency in the application of appropriate boundary conditions, need for a thermomechanical treatment of the heat generation in the dynamic loading condition, and further difficulties in model calibration based on limited real-world engineering data. As with the prior challenge, this work not only documents the 'state-of-the-art' in computational failure prediction of ductile tearing scenarios, but also provides a detailed dataset for non-blind assessment of alternative methods.
In recent years, substantial progress has been made in the field of reverberant speech signal processing, including both single- and multichannel dereverberation techniques and automatic speech recognition (ASR) techniques that are robust to reverberation. In this paper, we describe the REVERB challenge, which is an evaluation campaign that was designed to evaluate such speech enhancement (SE) and ASR techniques to reveal the state-of-the-art techniques and obtain new insights regarding potential future research directions. Even though most existing benchmark tasks and challenges for distant speech processing focus on the noise robustness issue and sometimes only on a single- channel scenario, a particular novelty of the REVERB challenge is that it is carefully designed to test robustness against reverberation, based on both real, single- channel, and multichannel recordings. This challenge attracted 27 papers, which represent 25 systems specifically designed for SE purposes and 49 systems specifically designed for ASR purposes. This paper describes the problems dealt within the challenge, provides an overview of the submitted systems, and scrutinizes them to clarify what current processing strategies appear effective in reverberant speech processing.
Silicon micro venturi nozzles for cost-efficient spray coating of thin organic P3HT/PCBM layers
(2017)
Improvements on spray coating are of particular interest to different fields of technology as it is a scalable deposition method and processing from solutions offer various application possibilities outside of typical facilities. When it comes to the deposition of expensive and film-forming media such as organic semiconductors, consumption and nozzle cleaning issues are of particular importance. We demonstrate the simple steps to design and fabricate micro venturi nozzles for economical spray coating with a consumption as low as 30-50 mu l.min(-1). For spray coating an active area of 25 cm(2) a 2.45-4.01 fold coating efficiency is observed compared to a conventional airbrush nozzle set. The electrical characterization of first diodes sprayed with an active layer thickness of similar to 750 nm using a single micronozzle at a coating speed of 1.7 cm(2).min(-1) reveals a good external quantum efficiency of 72.9% at 532 nm and a dark current of similar to 7.4.10(-5) mA.cm(-2), both measured at -2 V. Furthermore, the high resistance of the micronozzles against solvents and most acids is provided through realization in a silicon wafer with silicon dioxide encapsulation, therefore allowing easy and effective cleaning.
In this paper, we consider the reconstruction problem of photoacoustic tomography (PAT) with a flat observation surface. We develop a direct reconstruction method that employs regularization with wavelet sparsity constraints. To that end, we derive a wavelet-vaguelette decomposition (WVD) for the PAT forward operator and a corresponding explicit reconstruction formula in the case of exact data. In the case of noisy data, we combine the WVD reconstruction formula with soft-thresholding, which yields a spatially adaptive estimation method. We demonstrate that our method is statistically optimal for white random noise if the unknown function is assumed to lie in any Besov-ball. We present generalizations of this approach and, in particular, we discuss the combination of PAT-vaguelette soft-thresholding with a total variation (TV) prior. We also provide an efficient implementation of the PAT-vaguelette transform that leads to fast image reconstruction algorithms supported by numerical results.
Purpose
Age-related macular degeneration (AMD) is a common threat to vision. While classification of disease stages is critical to understanding disease risk and progression, several systems based on color fundus photographs are known. Most of these require in-depth and time-consuming analysis of fundus images. Herein, we present an automated computer-based classification algorithm.
Design Algorithm development for AMD classification based on a large collection of color fundus images. Validation is performed on a cross-sectional, population-based study.
Participants.
We included 120 656 manually graded color fundus images from 3654 Age-Related Eye Disease Study (AREDS) participants. AREDS participants were >55 years of age, and non-AMD sight-threatening diseases were excluded at recruitment. In addition, performance of our algorithm was evaluated in 5555 fundus images from the population-based Kooperative Gesundheitsforschung in der Region Augsburg (KORA; Cooperative Health Research in the Region of Augsburg) study.
Methods.
We defined 13 classes (9 AREDS steps, 3 late AMD stages, and 1 for ungradable images) and trained several convolution deep learning architectures. An ensemble of network architectures improved prediction accuracy. An independent dataset was used to evaluate the performance of our algorithm in a population-based study.
Main Outcome Measures.
κ Statistics and accuracy to evaluate the concordance between predicted and expert human grader classification.
Results.
A network ensemble of 6 different neural net architectures predicted the 13 classes in the AREDS test set with a quadratic weighted κ of 92% (95% confidence interval, 89%–92%) and an overall accuracy of 63.3%. In the independent KORA dataset, images wrongly classified as AMD were mainly the result of a macular reflex observed in young individuals. By restricting the KORA analysis to individuals >55 years of age and prior exclusion of other retinopathies, the weighted and unweighted κ increased to 50% and 63%, respectively. Importantly, the algorithm detected 84.2% of all fundus images with definite signs of early or late AMD. Overall, 94.3% of healthy fundus images were classified correctly.
Conclusions
Our deep learning algoritm revealed a weighted κ outperforming human graders in the AREDS study and is suitable to classify AMD fundus images in other datasets using individuals >55 years of age.