Filtern
Dokumenttyp
Sprache
- Englisch (11)
Schlagworte
- SAXS (4)
- Laser thermography (3)
- Nanoparticle (3)
- Neural network (3)
- Small-angle X-ray scattering (3)
- Active thermal imaging (2)
- Deep learning (2)
- Deep unfolding (2)
- Defect reconstruction (2)
- Elastic net (2)
Organisationseinheit der BAM
This article presents deep unfolding neural networks to handle inverse problems in photothermal radiometry enabling super-resolution (SR) imaging. The photothermal SR approach is a well-known technique to overcome the spatial resolution limitation in photothermal imaging by extracting high-frequency spatial components based on the deconvolution with the thermal point spread function (PSF). However, stable deconvolution can only be achieved by using the sparse structure of defect patterns, which often requires tedious, handcrafted tuning of hyperparameters and results in computationally intensive algorithms. On this account, this article proposes Photothermal-SR-Net, which performs deconvolution by deep unfolding considering the underlying physics. Since defects appear sparsely in materials, our approach includes trained block-sparsity thresholding in each convolutional layer. This enables to super-resolve 2-D thermal images for nondestructive testing (NDT) with a substantially improved convergence rate compared to classic approaches. The performance of the proposed approach is evaluated on various deep unfolding and thresholding approaches. Furthermore, we explored how to increase the reconstruction quality and the computational performance. Thereby, it was found that the computing time for creating high-resolution images could be significantly reduced without decreasing the reconstruction quality by using pixel binning as a preprocessing step.
Learned block iterative shrinkage thresholding algorithm for photothermal super resolution imaging
(2020)
Block-sparse regularization is already well-known in active thermal imaging and is used for multiple measurement based inverse problems. The main bottleneck of this method is the choice of regularization parameters which differs for each experiment. To avoid time-consuming manually selected regularization parameters, we propose a learned block-sparse optimization approach using an iterative algorithm unfolded into a deep neural network. More precisely, we show the benefits of using a learned block iterative shrinkage thresholding algorithm that is able to learn the choice of regularization parameters. In addition, this algorithm enables the determination of a suitable weight matrix to solve the underlying inverse problem. Therefore, in this paper we present the algorithm and compare it with state of the art block iterative shrinkage thresholding using synthetically generated test data and experimental test data from active thermography for defect reconstruction. Our results show that the use of the learned block-sparse optimization approach provides smaller normalized mean square errors for a small fixed number of iterations than without learning. Thus, this new approach allows to improve the convergence speed and only needs a few iterations to generate accurate defect reconstruction in photothermal super resolution imaging.
This paper presents deep unfolding neural networks to handle inverse problems in photothermal radiometry enabling super resolution (SR) imaging. Photothermal imaging is a well-known technique in active thermography for nondestructive inspection of defects in materials such as metals or composites. A grand challenge of active thermography is to overcome the spatial resolution limitation imposed by heat diffusion in order to accurately resolve each defect. The photothermal SR approach enables to extract high-frequency spatial components based on the deconvolution with the thermal point spread function. However, stable deconvolution can only be achieved by using the sparse structure of defect patterns, which often requires tedious, hand-crafted tuning of hyperparameters and results in computationally intensive algorithms. On this account, Photothermal-SR-Net is proposed in this paper, which performs deconvolution by deep unfolding considering the underlying physics. This enables to super resolve 2D thermal images for nondestructive testing with a substantially improved convergence rate. Since defects appear sparsely in materials, Photothermal-SR-Net applies trained blocksparsity thresholding to the acquired thermal images in each convolutional layer. The performance of the proposed approach is evaluated and discussed using various deep unfolding and thresholding approaches applied to 2D thermal images. Subsequently, studies are conducted on how to increase the reconstruction quality and the computational performance of Photothermal-SR-Net is evaluated.
Thereby, it was found that the computing time for creating high-resolution images could be significantly reduced without decreasing the reconstruction quality by using pixel binning as a preprocessing step.
Learned block iterative shrinkage thresholding algorithm for photothermal super resolution imaging
(2022)
Block-sparse regularization is already well known in active thermal imaging and is used for multiple-measurement-based inverse problems. The main bottleneck of this method is the choice of regularization parameters which differs for each experiment. We show the benefits of using a learned block iterative shrinkage thresholding algorithm (LBISTA) that is able to learn the choice of regularization parameters, without the need to manually select them. In addition, LBISTA enables the determination of a suitable weight matrix to solve the underlying inverse problem. Therefore, in this paper we present LBISTA and compare it with state-of-the-art block iterative shrinkage thresholding using synthetically generated and experimental test data from active thermography for defect reconstruction. Our results show that the use of the learned block-sparse optimization approach provides smaller normalized mean square errors for a small fixed number of iterations. Thus, this allows us to improve the convergence speed and only needs a few iterations to generate accurate defect reconstruction in photothermal super-resolution imaging.
The application of appropriate analytical techniques is essential for nanomaterial (NM) characterization. In this study, we compared different analytical techniques for NM analysis. Regarding possible adverse health effects, ionic and particulate NM effects have to be taken into account. As NMs behave quite differently in physiological media, special attention was paid to techniques which are able to determine the biosolubility and complexation behavior of NMs. Representative NMs of similar size were selected: aluminum (Al0) and aluminum oxide (Al2O3), to compare the behavior of metal and metal oxides. In addition, titanium dioxide (TiO2) was investigated. Characterization techniques such as dynamic light scattering (DLS) and nanoparticle tracking analysis (NTA) were evaluated with respect to their suitability for fast characterization of nanoparticle dispersions regarding a particle's hydrodynamic diameter and size distribution. By application of inductively coupled plasma mass spectrometry in the single particle mode (SP-ICP-MS), individual nanoparticles were quantified and characterized regarding their size. SP-ICP-MS measurements were correlated with the information gained using other characterization techniques, i.e. transmission electron microscopy (TEM) and small angle X-ray scattering (SAXS). The particle surface as an important descriptor of NMs was analyzed by X-ray diffraction (XRD). NM impurities and their co-localization with biomolecules were determined by ion beam microscopy (IBM) and confocal Raman microscopy (CRM). We conclude advantages and disadvantages of the different techniques applied and suggest options for their complementation. Thus, this paper may serve as a practical guide to particle characterization techniques.
Aluminum (Al) can be taken up from food, packaging, or the environment and thus reaches the human gastrointestinal tract. Its toxic potential after oral uptake is still discussed. The fate of different solid and ionic Al species during the passage through the digestive tract is the focus of this research, as well as the cellular effects caused by these different Al species. The present study combines the physicochemical processing of three recently studied Al species (metallic Al0, mineral Al2O3, and soluble AlCl3) in artificial digestion fluids with in vitro cell systems for the human intestinal barrier. Inductively coupled plasma mass spectrometry (ICP-MS) and small-angle X-ray scattering (SAXS) methods were used to characterize the Al species in the artificial digestion fluids and in cell culture medium for proliferating and differentiated intestinal Caco-2 cells. Cytotoxicity testing and cellular impedance measurements were applied to address the effects of digested Al species on cell viability and cell proliferation. Microarray-based transcriptome analyses and quantitative real-time PCR were conducted to obtain a deeper insight into cellular mechanisms of action and generated indications for cellular oxidative stress and an influence on xenobiotic metabolism, connected with alterations in associated signaling pathways. These cellular responses, which were predominantly caused by formerly ionic Al species and only at very high concentrations, were not impacted by artificial digestion. A two-directional conversion of Al between ionic species and solid particles occurred throughout all segments of the gastrointestinal tract, as evidenced by the presence of nanoscaled particles. Nevertheless, this presence did not increase the toxicity of the respective Al species.
Classification of Spot-Welded Joints in Laser Thermography Data Using Convolutional Neural Networks
(2021)
Spot welding is a crucial process step in various industries. However, classification of spot welding quality is still a tedious process due to the complexity and sensitivity of the test material, which drain conventional approaches to its limits. In this article, we propose an approach for quality inspection of spot weldings using images from laser thermography data. We propose data preparation approaches based on the underlying physics of spot-welded joints, heated with pulsed laser thermography by analyzing the intensity over time and derive dedicated data filters to generate training datasets. Subsequently, we utilize convolutional neural networks to classify weld quality and compare the performance of different models against each other. We achieve competitive results in terms of classifying the different welding quality classes compared to traditional approaches, reaching an accuracy of more than 95 percent. Finally, we explore the effect of different augmentation methods.
Uptake and molecular impact of aluminum-containing nanomaterials on human intestinal caco-2 cells
(2018)
Aluminum (Al) is one of the most common elements in the earth crust and increasingly used in food, consumer products and packaging. Its hazard potential for humans is still not completely understood. Besides the metallic form, Al also exists as mineral, including the insoluble oxide, and in soluble ionic forms. Representatives of these three species, namely a metallic and an oxidic species of Al-containing nanoparticles and soluble aluminum chloride, were applied to human intestinal cell lines as models for the intestinal barrier. We characterized physicochemical particle parameters, protein corona composition, ion release and cellular uptake. Different in vitro assays were performed to determine potential effects and molecular modes of Action related to the individual chemical species. For a deeper insight into signaling processes, microarray transcriptome analyses followed by bioinformatic data analysis were employed. The particulate Al species showed different solubility in biological media. Metallic Al nanoparticles released more ions than Al2O3 nanoparticles, while AlCl3 showed a mixture of dissolved and agglomerated particulate entities in biological media. The protein corona composition differed between both nanoparticle species. Cellular uptake, investigated in transwell experiments, occurred predominantly in particulate form, whereas ionic Al was not taken up by intestinal cell lines. Transcellular transport was not observed. None of the Al species showed cytotoxic effects up to 200 mg Al/mL. The transcriptome analysis indicated mainly effects on oxidative stress pathways, xenobiotic metabolism and metal homeostasis. We have shown for the first time that intestinal cellular uptake of Al occurs preferably in the particle form, while toxicological effects appear to be ion-related.
Aluminum has gathered toxicological Attention based on relevant human exposure and its suspected hazardous potential. Nanoparticles from food supplements or Food contact materials may reach the human gastrointestinal tract.
Here, we monitored the physicochemical fate of aluminum containing nanoparticles and aluminum ions when passaging an in vitro model of the human gastrointestinal tract. Smallangle X-ray scattering (SAXS), transmission electron microscopy (TEM), ion beam microscopy (IBM), secondary ion beam mass spectrometry (TOF-SIMS), and inductively coupled plasma mass spectrometry (ICP-MS) in the singleparticle mode were employed to characterize two aluminumcontaining nanomaterials with different particle core materials (Al0, γAl2O3) and soluble AlCl3. Particle size and shape remained unchanged in saliva, whereas strong Agglomeration of both aluminum nanoparticle species was observed at low pH in gastric fluid together with an increased ion release. The levels of free aluminum ions decreased in intestinal fluid and the particles deagglomerated, thus liberating primary particles again. Dissolution of nanoparticles was limited and substantial changes of their shape and size were not detected. The amounts of particle-associated phosphorus, chlorine, potassium, and calcium increased in intestinal fluid, as compared to nanoparticles in standard dispersion.
Interestingly, nanoparticles were found in the intestinal fluid after addition of ionic aluminum. We provide a comprehensive characterization of the fate of aluminum nanoparticles in simulated gastrointestinal fluids, demonstrating that orally ingested nanoparticles probably reach the intestinal epithelium. The balance between dissolution and de novo complex formation should be considered when evaluating nanotoxicological experiments.
Size and shape are crucial parameters which have impact on the potential of nanoparticles to penetrate cell membranes and epithelial barriers. Current research in nanotoxicology additionally focuses on particle coating. To distinguish between core- and coating-related effects in nanoparticle uptake and translocation, two nanoparticles equal in size, coating and charge but different in core material were investigated.
Silver and iron oxide nanoparticles coated with poly(acrylic acid) were chosen and extensively characterized by small-angle x-ray scattering, nanoparticle tracing analysis and transmission electron microscopy (TEM). Uptake and transport were studied in the intestinal Caco-2 model in a Transwell System with subsequent elemental analysis. TEM and ion beam microscopy were conducted for particle visualization.
Although equal in size, charge and coating, the behavior of the two particles in Caco-2 cells was different: while the internalized amount was comparable, only iron oxide nanoparticles additionally passed the epithelium. Our findings suggest that the coating material influenced only the uptake of the nanoparticles whereas the translocation was determined by the core material.
Knowledge about the different roles of the particle coating and core materials in crossing biological barriers will facilitate toxicological risk assessment of nanoparticles and contribute to the optimization of pharmacokinetic properties of nano-scaled pharmaceuticals.