SynDistNet: self-supervised monocular fisheye camera distance estimation synergized with semantic segmentation for autonomous driving. - In: 2021 IEEE Winter Conference on Applications of Computer Vision, (2021), S. 61-71
State-of-the-art self-supervised learning approaches for monocular depth estimation usually suffer from scale ambiguity. They do not generalize well when applied on distance estimation for complex projection models such as in fisheye and omnidirectional cameras. This paper introduces a novel multi-task learning strategy to improve self-supervised monocular distance estimation on fisheye and pinhole camera images. Our contribution to this work is threefold: Firstly, we introduce a novel distance estimation network architecture using a self-attention based encoder coupled with robust semantic feature guidance to the decoder that can be trained in a one-stage fashion. Secondly, we integrate a generalized robust loss function, which improves performance significantly while removing the need for hyperparameter tuning with the reprojection loss. Finally, we reduce the artifacts caused by dynamic objects violating static world assumptions using a semantic masking strategy. We significantly improve upon the RMSE of previous work on fisheye by 25% reduction in RMSE. As there is little work on fisheye cameras, we evaluated the proposed method on KITTI using a pinhole model. We achieved state-of-the-art performance among self-supervised methods without requiring an external scale estimation.
https://doi.org/10.1109/WACV48630.2021.00011
Virtual environments for Smart House system studying. - In: Educating engineers for future industrial revolutions, (2021), S. 569-576
Today, virtual worlds are used not only in the gaming industry. Virtual, augmented and cross-reality also allow to organize effectively 3D environments that provide the effect of immersion and user interaction with objects and processes of the learning environment. The combination of physical objects and virtual models is a modern trend in the online lab development. Based on this approach, online experiment becomes more visual and interesting for students. In addition, it reduces queues for experiments on the real equipment. The paper presents the results of the development of an interactive web-oriented virtual model that expands the Smart House & IoT remote laboratory functionality as well as 3D virtual environment of Smart House for the virtual reality helmet. The proposed solutions will motivate students to study home automation technologies and the features of Smart Home systems realization.
Modern mobile interface for remote laboratory control. - In: Educating engineers for future industrial revolutions, (2021), S. 584-592
The implementation of modern mobile interfaces for online laboratories is an urgent task because it allows increasing students' interest in such educational resources usage, especially in non-standard situations (for example, during self-isolation period). The results of Telegram messenger's chatbot development for interaction with remote laboratory Smart House & IoT are presented in this paper. It gives comfortable and effective tools and possibilities for the popularization of remote laboratory application for home automation technologies studying.
Explaining clinical decision support systems in medical imaging using cycle-consistent activation maximization. - In: Neurocomputing, ISSN 1872-8286, Bd. 458 (2021), S. 141-156
Clinical decision support using deep neural networks has become a topic of steadily growing interest. While recent work has repeatedly demonstrated that deep learning offers major advantages for medical image classification over traditional methods, clinicians are often hesitant to adopt the technology because its underlying decision-making process is considered to be intransparent and difficult to comprehend. In recent years, this has been addressed by a variety of approaches that have successfully contributed to providing deeper insight. Most notably, additive feature attribution methods are able to propagate decisions back into the input space by creating a saliency map which allows the practitioner to "see what the network sees." However, the quality of the generated maps can become poor and the images noisy if only limited data is available - a typical scenario in clinical contexts. We propose a novel decision explanation scheme based on CycleGAN activation maximization which generates high-quality visualizations of classifier decisions even in smaller data sets. We conducted a user study in which we evaluated our method on the LIDC dataset for lung lesion malignancy classification, the BreastMNIST dataset for ultrasound image breast cancer detection, as well as two subsets of the CIFAR-10 dataset for RBG image object recognition. Within this user study, our method clearly outperformed existing approaches on the medical imaging datasets and ranked second in the natural image setting. With our approach we make a significant contribution towards a better understanding of clinical decision support systems based on deep neural networks and thus aim to foster overall clinical acceptance.
https://doi.org/10.1016/j.neucom.2021.05.081
A generative approach for creating Eclipse Sirius editors for generic systems. - In: SYSCON 2021, (2021), insges. 8 S.
Model-Driven Engineering (MDE) is getting more and more important for modeling, analyzing, and simulating complicated systems. It can also be used for both documenting and generating source code, which is less error-prone than a manually written one. For defining a model, it is common to have a graphical representation that can be edited through an editor. Creating such an editor for a given domain may be a difficult task for first-time users and a tedious, repetitive, and error-prone task for experienced ones. This paper introduces a new automated flow to ease the creation of ready-to-use Sirius editors based on a model, graphically defined by the domain experts, which describe their domains structure. We provide different model transformations to generate the required artifacts to obtain a fully-fledged Sirius editor based on a generated domain metamodel. The generated editor can then be distributed as an Eclipse application or as a collaborative web application. Thanks to this generative approach, it is possible to reduce the cost of refactoring the domains model in successive iterations, as only the final models need to be updated to conform to the latest format. At the same time, the editor gets generated and hence updated automatically at practically no cost.
https://doi.org/10.1109/SysCon48628.2021.9447062
Surface-modified magnetite nanoparticles affect lysozyme amyloid fibrillization. - In: Biochimica et biophysica acta, ISSN 1872-8006, Bd. 1865 (2021), 9, 129941, insges. 9 S.
Background - The surface of nanoparticles (NPs) is an important factor affecting the process of poly/peptides' amyloid aggregation. We have investigated the in vitro effect of trisodium citrate (TC), gum arabic (GA) and citric acid (CA) surface-modified magnetite nanoparticles (COAT-MNPs) on hen egg-white lysozyme (HEWL) amyloid fibrillization and mature HEWL fibrils. - Methods - Dynamic light scattering (DLS) was used to characterize the physico-chemical properties of studied COAT-MNPs and determine the adsorption potential of their surface towards HEWL. The anti-amyloid properties were studied using thioflavin T (ThT) and tryptophan (Trp) intrinsic fluorescence assays, and atomic force microscopy (AFM). The morphology of amyloid aggregates was analyzed using Gwyddion software. The cytotoxicity of COAT-MNPs was determined utilizing Trypan blue (TB) assay. - Results - Agents used for surface modification affect the COAT-MNPs physico-chemical properties and modulate their anti-amyloid potential. The results from ThT and intrinsic fluorescence showed that the inhibitory activities result from the more favorable interactions of COAT-MNPs with early pre-amyloid species, presumably reducing nuclei and oligomers formation necessary for amyloid fibrillization. COAT-MNPs also possess destroying potential, which is presumably caused by the interaction with hydrophobic residues of the fibrils, resulting in the interruption of an interface between [beta]-sheets stabilizing the amyloid fibrils. - Conclusion - COAT-MNPs were able to inhibit HEWL fibrillization and destroy mature fibrils with different efficacy depending on their properties, TC-MNPs being the most potent nanoparticles. - General significance - The study reports findings regarding the general impact of nanoparticles' surface modifications on the amyloid aggregation of proteins.
https://doi.org/10.1016/j.bbagen.2021.129941
Sparse modeling and monitoring for industrial processes using sparse, distributed principal component analysis. - In: Journal of the Taiwan Institute of Chemical Engineers, ISSN 1876-1089, Bd. 122 (2021), S. 14-22
Driven by the strong demand for sparsity in dimensional reduction techniques, a sparse modeling and monitoring approach based on sparse, distributed principal component analysis (SDPCA) is proposed to achieve sparsity. To this end, the data set is first divided into highly correlated blocks (HCBs) and one remainder block (RB) on the basis of the mutual-information-based correlation matrix. From this, the sparse loading vectors for the HCBs are obtained using the PCA models, while for the RB, it is obtained using the sparse PCA model. It is worth noting that the sparsity in SDPCA enables the sparse loading vectors to produce interpretable principal components, which keeps the correlations between the highly correlated variables and achieves the sparsity for the weakly correlated ones. Moreover, to fully appreciate the interpretation of the sparse principal components, a fault diagnosis strategy named blockwise contribution plots is proposed by first determining the faulty block, and then, identifying the faulty variables. Compared with PCA and SPCA, the proposed SDPCA detects more faulty samples and gives more accurate diagnosis results.
https://doi.org/10.1016/j.jtice.2021.04.029
Mobile robot-based gait training after total hip arthroplasty (THA) improves walking in biomechanical gait analysis. - In: Journal of Clinical Medicine, ISSN 2077-0383, Bd. 10 (2021), 11, 2416, insges. 11 S.
There are multiple attempts to decrease costs in the healthcare system while maintaining a high treatment quality. Digital therapies receive increasing attention in clinical practice, mainly relating to home-based exercises supported by mobile devices, eventually in combination with wearable sensors. The aim of this study was to determine if patients following total hip arthroplasty (THA) could benefit from gait training on crutches conducted by a mobile robot in a clinical setting. Method: This clinical trial was conducted with 30 patients following total hip arthroplasty. Fifteen patients received the conventional physiotherapy program in the clinic (including 5 min of gait training supported by a physiotherapist). The intervention group of 15 patients passed the same standard physiotherapy program, but the 5-min gait training supported by a physiotherapist was replaced by 2 × 5 min of gait training conducted by the robot. Length of stay of the patients was set to five days. Biomechanical gait parameters of the patients were assessed pre-surgery and upon patient discharge. Results: While before surgery no significant difference in gait parameters was existent, patients from the intervention group showed a significant higher absolute walking speed (0.83 vs. 0.65 m/s, p = 0.029), higher relative walking speed (0.2 vs. 0.16 m/s, p = 0.043) or shorter relative cycle time (3.35 vs. 3.68 s, p = 0.041) than the patients from the control group. Conclusion: The significant higher walking speed of patients indicates that such robot-based gait training on crutches may shorten length of stay (LOS) in acute clinics. However, the number of patients involved was rather small, thus calling for further studies.
https://doi.org/10.3390/jcm10112416
Adaptive query compilation in graph databases. - In: 2021 IEEE 37th International Conference on Data Engineering workshops, (2021), S. 112-119
Compiling database queries into compact and efficient machine code has proven to be a great technique to improve query performance and to exploit characteristics of modern hardware. Furthermore, compilation frameworks like LLVM provide powerful optimization techniques and support different backends. However, the time for generating machine code becomes an issue for short-running queries or queries which could produce early results quickly. In this work, we present an adaptive approach integrating graph query interpretation and compilation. While query compilation and code generation are running in the background, the query execution starts using the interpreter. As soon as the code generation is finished, the execution switches to the compiled code. Our evaluation shows that autonomously switching execution modes helps to hide compilation times.
https://doi.org/10.1109/ICDEW53142.2021.00027
T2- and T1 relaxivities and magnetic hyperthermia of iron-oxide nanoparticles combined with paramagnetic Gd complexes. - In: Journal of chemical sciences, ISSN 0973-7103, Bd. 133 (2021), 2, 43, insges. 10 S.
Im Titel sind "2-" und "1" tiefgestellt
The present paper reports the synthesis of iron-oxide nanoparticles (diameter 12.8±2.2 nm) coated with silica shell doped with paramagnetic Gd(III)-based complexes. The resulting nanoparticles with a silica shell thickness of about 45 nm have an average diameter of 113.1±14.3 nm and feature high transverse and longitudinal relaxivities (356 and 25 mM^-1 s^-1, respectively) at 1.5 T and 25 ˚C on a medical whole body NMR scanner. It has been also revealed using magnetic heating measurements that the prepared core-shell nanoparticles possess a high specific adsorption rate of around 236 W/g in aqueous media. The surface of the composite nanoparticles was decorated by amino-groups for a greater cellular uptake behaviour. The cell viability measurements reveal the concentration-dependent cytotoxicity of the nanoparticles, which agrees well with the high content of Gd(III) complexes in the nanomaterial. The obtained results show that the core-shell design of nanoparticles with superparamagnetic and paramagnetic parts can be promising for high transverse (and longitudinal) relaxivity as well as magnetic hyperthermia.
https://doi.org/10.1007/s12039-021-01904-7