Regensburg Center of Health Sciences and Technology - RCHST
Refine
Year of publication
- 2021 (5) (remove)
Document Type
- conference proceeding (article) (5) (remove)
Has Fulltext
- no (5)
Is part of the Bibliography
- no (5)
Keywords
- Deep Learning (2)
- Computerunterstützte Medizin (1)
- Convolutional neural networks (1)
- Deep learning fluid mechanics (1)
- Endoskopie (1)
- Multiphase blood flow (1)
- Non-Newtonian multiphase flow (1)
- Patient (1)
- Roboter (1)
- Schlaganfall (1)
Institute
- Regensburg Center of Health Sciences and Technology - RCHST (5)
- Fakultät Maschinenbau (3)
- Regensburg Center of Biomedical Engineering - RCBE (3)
- Fakultät Informatik und Mathematik (2)
- Labor Biofluidmechanik (2)
- Fakultät Angewandte Sozial- und Gesundheitswissenschaften (1)
- Labor Biomechanik (LBM) (1)
- Labor Empirische Sozialforschung (1)
- Labor Logopädie (LP) (1)
- Labor Pflegeforschung (LPF) (1)
Begutachtungsstatus
- peer-reviewed (4)
Convolutional Neural Networks for Approximation of Internal Non-Newtonian Multiphase Flow Fields
(2021)
Neural networks (NNs) as an alternative method for universal approximation of differential equations have proven to be computationally efficient and still sufficiently accurate compared to established methods such as the finite volume method (FVM). Additionally, analysing weights and biases can give insights into the underlying physical laws. FVM and NNs are both based upon spacial discretisation. Since a Cartesian and equidistant grid is a raster graphics, image-to-image regression techniques can be used to predict phase velocity fields as well as particle and pressure distributions from simple mass flow boundary conditions. The impact of convolution layer depth and number of channels of a ConvolutionDeconvolution Regression Network (CDRN), on prediction performance of internal non-Newtownian multiphase flows is investigated. Parametric training data with 2055 sets is computed using FVM. To capture significant non-Newtownian effects of a particle-laden fluid (e.g. blood) flowing through small and non-straight channels, an Euler-Euler multiphase approach is used. The FVM results are normalized and mapped onto an equidistant grid as supervised learning target. The investigated NNs consist of n= {3, 5, 7} corresponding encoding/decoding blocks and different skip connections. Regardless of the convolution depth (i.e. number of blocks), the deepest spacial down-sampling via strided convolution is adjusted to result in a 1 × 1 × f · 2nfeature map, with f = {8, 16, 32}. The prediction performance expressed is as channel-averaged normalized root mean squared error (NRMSE). With a NRMSE of < 2 · 10-3, the best preforming NN has f = 32 initial feature maps, a kernel size of k = 4, n = 5 blocks and dense skip connections. Average inference time from this NN takes < 7 · 10-3s. Worst accuracy at NRMSE of approx 9 · 10-3is achieved without any skips, at k = 2, f = 16 and n = 3, but deployment takes only < 2 · 10-3s Given an adequate training, the prediction accuracy improves with convolution depth, where more features have higher impact on deeper NNs. Due to skip connections and batch normalisation, training is similarly efficient, regardless of the depth. This is further improved by blocks with dense connections, but at the price of a drastically larger model. Depending on geometrical complexity, spacial resolution is critical, as it increases the number of learnables and memory requirements massively.
Blood flow in channels of varying diameters <500μm exhibits strong non-linear effects. Multiphase finite volume approaches are feasible, but still computationally costly. Here, the feasibility of applying convolutional neural networks for blood flow prediction in artificial lungs is investigated. Training targets are precomputed using an Eulerian two-phase approach. To match with experimental data, the interphase drag and lift, as well as intraphase shear-thinning are adapted. A recursively branching regression network and convolution/deconvolution networks with plain skip connections and densely connected skips are investigated. A priori knowledge is incorporated in the loss functional to prevent the network from learning non-physical solutions. Inference from neural networks is approximately six orders of magnitude faster than the classical finite volume approach. Even if resulting in comparably coarse flow fields, the neural network predictions can be used as close to convergence initial solutions greatly accelerating classical flow computations.
Ergonomic workplaces lead to fewer work-related musculoskeletal disorders and thus fewer sick days. There are various guidelines to help avoid harmful situations. However, these recommendations are often rather crude and often neglect the complex interaction of biomechanical loading and psychological stress. This study investigates whether machine learning algorithms can be used to predict mechanical and stress-related muscle activity for a standardized motion. For this purpose, experimental data were collected for trunk movement with and without additional psychological stress. Two different algorithms (XGBoost and TensorFlow) were used to model the experimental data. XGBoost in particular predicted the results very well. By combining it with musculoskeletal models, the method shown here can be used for workplace analysis but also for the development of real-time feedback systems in real workplace environments.
Vorgestellt wird das Forschungsvorhaben des interdisziplinären Forschungsprojekts „DeinHaus 4.0 – Telepräsenzroboter für die Pflege und Unterstützung von Schlaganfallpatientinnen und -patienten (TePUS)“. Über geeignete Robotersysteme sollen den Proband*innen individuell abgestimmte Angebote aus den Bereichen Telenursing und Teletherapie
zur Verfügung gestellt werden. Hierbei werden einerseits Telepräsenzsitzungen durchgeführt, anderseits kommen verschiedene Apps zum Einsatz. Die Interventionen werden mittels eines Mixed-Methods-Ansatzes evaluiert. Begleitend werden Technikakzeptanz und Nutzungsbereitschaft untersucht.
Barrett's esophagus denotes a disorder in the digestive system that affects the esophagus' mucosal cells, causing reflux, and showing potential convergence to esophageal adenocarcinoma if not treated in initial stages. Thus, fast and reliable computer-aided diagnosis becomes considerably welcome. Nevertheless, such approaches usually suffer from imbalanced datasets, which can be addressed through Generative Adversarial Networks (GANs). Such techniques generate realistic images based on observed samples, even though at the cost of a proper selection of its hyperparameters. Many works employed a class of nature-inspired algorithms called metaheuristics to tackle the problem considering distinct deep learning approaches. Therefore, this paper's main contribution is to introduce metaheuristic techniques to fine-tune GANs in the context of Barrett's esophagus identification, as well as to investigate the feasibility of generating high-quality synthetic images for early-cancer assisted identification.