Search Results (33)

The present research deals with two possible sources of bias that arise naturally from the selection procedure when modeling extreme phenomena. More specifically, the first type of bias arises when an r-size-biased sample from a set of maximum values is selected, while the second one occurs when a random sample of maxima is observed where each observation is obtained by a series of r-size-biased samples. The concept of weighted distributions is used, not only to describe both cases but also as an adjustment methodology. The differences between the two types of bias are discussed, while the impact of ignoring the bias on the estimation of the unknown parameters is revealed both theoretically and with the use of a simulation study, under the assumption that the parent distribution belongs to the Fréchet maximum domain of attraction. Finally, numerical results indicate that ignorance of the bias or misspecification of r results in inconsistent estimators. Full article

In this work, we describe a new type of electrochemical aptasensor for the label-free detection of oxytetracycline (OTC). Thin-film gold electrodes were fabricated through sputtering gold on a Kapton film, followed by the immobilization of a thiol-modified aptamer on the electrode surface. The selective capture of OTC at the aptamer-functionalized electrodes was monitored electrochemically with the use of the [Fe(CN)₆]⁴⁻/[Fe(CN)₆]³⁻ redox probe. Different experimental variables were studied, through which the metrological features for OTC determination were derived. Finally, the developed sensor was implemented to achieve the detection of OTC in a spiked milk sample. Full article

In open pit mining areas, knowledge of geotechnical conditions (e.g., overburden thickness, background slope, and fault locations) ensures geotechnical safety during exploitation as well as reclamation planning. The Greek Public Power Corporation initiated a research program after stability issues emerged on the southern side of the Mavropigi open pit mine in NW Macedonia. Geotechnical wells revealed steeply dipping bedrock and thin tectonic contact, indicating the need for the detailed imaging of the subsurface for future stability measures. For this purpose, a geophysical investigation aimed to extract information mostly for the dip of the interface between schist bedrock and overlaying Neogene sediments and/or limestones. Based on the high contrast of electrical properties between schists and limestones, as well as the differences in acoustic impedance and formation thickness, the seismic reflection and electrical resistivity tomography (ERT) methods were selected. The suitability of the seismic reflection for its application in this area was checked by generating synthetic seismic data, which resulted from the simulation of seismic wave propagation for geological models of the area. The acquisition parameters were determined after the noise test. Field seismic data processing produced a depth-migrated section, which revealed the existence of a fault. The use of dipole–dipole and gradient arrays, in 2D and 3D electrical resistivity measurements, ensured both the lateral and vertical mapping of schist bedrock and detected limestone bodies within the overburden. Also, the tectonic contact zone between limestone and schist formations was properly imaged. The comparison between geoelectrical and seismic sections indicated that the seismic reflection method provided a more accurate estimate of fault inclination. Finally, the geophysical survey enriched the geotechnical models necessary for sustainable mining (e.g., rational exploitation, the optimization of productivity, and zero accidents) including the planning of future reclamation. Full article

This work introduces a methodology for the automatic unmineable inclusions detection and Bucket Wheel Excavator (BWE) collision prevention, using electromagnetic (EM) inspection and a fuzzy inference system. EM data are collected continuously ahead from the bucket wheel of a BWE and subjected to processing. Two distinct methodologies for data processing were developed and integrated into the MATLAB programming environment. The first approach, named “Simple Mode”, utilizes statistical process control to generate real-time alerts in the event of a potential collision involving the excavator’s bucket and hard rock inclusions. The advanced processing flow (“Advanced Mode”) requires accurate instrument positioning and data from successive EM scans. It incorporates techniques of local resistivity maxima detection (Position Prominence Index) as well as Neural Network-based Pattern Recognition (NNPR). A decision support process based on a Fuzzy Inference System (FIS) has been developed to assist BWE operators in avoiding collision when digging hard rock inclusions. The proposed methodology was extensively tested using synthetic EM data. Limited real data, acquired with a CMD2 (GF Instruments) EM instrument equipped with GPS, were used to control its efficiency. Increased accuracy in the automatic detection of unmineable inclusions was observed using the Advanced Mode. On the other hand, the Simple Mode processing technique offers the advantage of being independent of instrument positioning as well as it provides real-time inspection of the excavated mine slope. This work introduces a methodology for hard rock inclusion detection and can contribute to the optimization of mine operations by improving resource efficiency, safety, cost savings, and environmental sustainability. Full article

Poor air quality remains the largest environmental health risk in Europe, despite the EU policy efforts. Especially in cities, the synergistic interactions between the urban heat island and urban pollution result in premature mortality, associated with cardiovascular and respiratory diseases. Mediterranean urban areas are particularly susceptible under the consideration that the intensity, frequency, and duration of heat waves will increase due to climate change. The LIFE SIRIUS project designates that air quality management needs to go beyond traditional approaches in order to consider synergistic effects. This paper assesses the impact of temperature on daily mortality from 2004 to 2019 in the Republic of Cyprus with the use of a Generalized Additive Model (GAM). The association between mean daily temperature and mortality is nonlinear, implying that a prompt rise in deaths occurs when temperatures are high, while for colder temperatures, the effect is delayed. We report an inverted J-shaped relationship between mean temperature and mortality, with the most prominent effects on human health documented at low temperatures. The population under study appears to be acclimatized to local conditions, as mortality increases after 10 days of exposure to the environmental risk. The results of this study will assist in the definition of city-specific thresholds above which health warnings for the protection of the local population will be issued, in the framework of LIFE SIRIUS. Full article

Climate models are fundamental tools for assessing historical climate conditions and projecting future ones. However, the results often differ systematically from observational data. The minimization of these differences is known as bias correction. The present study aims to correct the biases between observed daily precipitation values and the respective simulated ones from a EURO-CORDEX climate model. For this purpose, powerful statistical tools—generalized additive models (GAMs)—are used. GAMs are modified to adjust the simulated rainfall with the highest accuracy, and subsequently, they are evaluated by comparison with observational data. The method was applied to two eastern Mediterranean stations (Larissa in Greece and Larnaca in Cyprus) for the period 1981 to 2005. The results from both stations reveal that GAMs offer a valuable and accurate technique for the bias adjustment of daily precipitation. Full article

Rosa damascena Mill. is commercially the most important rose species used to produce essential oils. The plants of this species, cultivated in the district of Western Macedonia (Greece) for rose oil production, originated from indigenous genotypes but also nurseries abroad, mainly from Bulgaria. The present study investigated the genetic relationship between nine genotypes of R. damascena from Greece, one genotype from Turkey, three genotypes from Bulgaria and three genotypes from France using the molecular markers ISSR and SCoT. Also, the rooting ability of shoot cuttings from these nine genotypes was investigated by applying 2 g/L of the rooting regulator K-IBA. In addition, petals were chemically analyzed using GC-MS and LC-MS to identify the compounds that are the main components of the rose oil. The nine rose genotypes of R. damascena, cultivated in Greece, one from Turkey and one of the three genotypes from Bulgaria were clustered in one clade in the dendrogram. The other two genotypes from Bulgaria were clustered in a separate clade that demonstrated the existence of genetic diversity among the three Bulgarian genotypes, while the French genotypes were clustered in a third clade. The shoot cuttings rooted relatively easily (55–70%) with the application of K-IBA, without any significant differences among the nine genotypes. Large variation was observed among the nine genotypes in the main volatile compounds of the flower petal extracts, which are related to rose oil components. For these compounds, the concentrations in μg/g of the fresh petal weight were 2-phenylethylalcohol (1148.35–2777.19), nerol (27.45–64.93), citronellol (88.45–206.59), geraniol (69.12–170.99) and nonadecane (209.27–533.15). Of the non-volatile compounds, gallic acid was the most abundant phenolic acid in the petal extracts of the nine genotypes (0.28–0.82 μg/g), while for the flavonoids, quercetin and kaempferol variations of 0.35–1.17 μg/g and 0.26–2.13 μg/g were recorded, respectively. Full article

The addition of biocarriers can improve biological processes in bioreactors, since their surface allows for the immobilization, attachment, protection, and growth of microorganisms. In addition, the development of a biofilm layer allows for the colonization of microorganisms in the biocarriers. The structure, composition, and roughness of the biocarriers’ surface are crucial factors that affect the development of the biofilm. In the current work, the aluminosilicate zeolites 13X and ZSM-5 were examined as the main building components of the biocarrier scaffolds, using bentonite, montmorillonite, and halloysite nanotubes as inorganic binders in various combinations. We utilized 3D printing to form pastes into monoliths that underwent heat treatment. The 3D-printed biocarriers were subjected to a mechanical analysis, including density, compression, and nanoindentation tests. Furthermore, the 3D-printed biocarriers were morphologically and structurally characterized using nitrogen adsorption at 77 K (LN₂), scanning electron microscopy (SEM), and X-ray diffraction (XRD). The stress–strain response of the materials was obtained through nanoindentation tests combined with the finite element analysis (FEA). These tests were also utilized to simulate the lattice geometries under compression loading conditions to investigate their deformation and stress distribution in relation to experimental compression testing. The results indicated that the 3D-printed biocarrier of 13X/halloysite nanotubes was endowed with a high specific surface area of 711 m²/g and extended mesoporous structure. Due to these assets, its bulk density of 1.67 g/cm³ was one of the lowest observed amongst the biocarriers derived from the various combinations of materials. The biocarriers based on the 13X zeolite exhibited the highest mechanical stability and appropriate morphological features. The 13X/halloysite nanotubes scaffold exhibited a hardness value of 45.64 MPa, which is moderate compared to the rest, while it presented the highest value of modulus of elasticity. In conclusion, aluminosilicate zeolites and their combinations with clays and inorganic nanotubes provide 3D-printed biocarriers with various textural and structural properties, which can be utilized to improve biological processes, while the most favorable characteristics are observed when utilizing the combination of 13X/halloysite nanotubes. Full article

Ammonia and sulfide derived from the reduction of sulfate by the sulfate-reducing bacteria (SRB) are two of the most common inhibitors in anaerobic digestion. Zeolites and bentonites are characterized as porous materials able to adsorb both ammonia and sulfur compounds and seem to be promising candidates as additives in anaerobic digestion to counteract this co-inhibition. In this study, bentonite and zeolite 13X were subjected to alkali modification at different concentrations of NaOH to alter their physicochemical properties, and their effect on the alleviation of ammonia and sulfate co-inhibition in anaerobic digestion of cow manure was examined. The methane production in 13X treatments (13X without NaOH, 13X02-NaOH 0.2 M and 13X1-NaOH 1 M) was elevated by increasing the NaOH concentration in the modification step, resulting in a significance increase by 8.96%, 11.0% and 15.56% in 13X treatments compared to the treatment without additive. Bentonite treatments did not show the same behavior on the toxicity mitigation. The results appear to be influenced by the combined effect of 13X zeolites on the sulfur compounds adsorption and on the increase in pH and Na⁺ concentration in the batch reactors. Full article

The recent advances in deep learning have revolutionised computer-aided diagnosis in medical imaging. However, deep learning approaches to unveil their full potential require significant amounts of data, which can be a challenging task in some scientific fields, such as musculoskeletal ultrasound imaging, in which data privacy and security reasons can lead to important limitations in the acquisition and the distribution process of patients’ data. For this reason, different generative methods have been introduced to significantly reduce the required amount of real data by generating synthetic images, almost indistinguishable from the real ones. In this study, the power of the diffusion models is incorporated for the generation of realistic data from a small set of musculoskeletal ultrasound images in four different muscles. Afterwards, the similarity of the generated and real images is assessed with different types of qualitative and quantitative metrics that correspond well with human judgement. In particular, the histograms of pixel intensities of the two sets of images have demonstrated that the two distributions are statistically similar. Additionally, the well-established LPIPS, SSIM, FID, and PSNR metrics have been used to quantify the similarity of these sets of images. The two sets of images have achieved extremely high similarity scores in all these metrics. Subsequently, high-level features are extracted from the two types of images and visualized in a two-dimensional space for inspection of their structure and to identify patterns. From this representation, the two sets of images are hard to distinguish. Finally, we perform a series of experiments to assess the impact of the generated data for training a highly efficient Attention-UNet for the important clinical application of muscle thickness measurement. Our results depict that the synthetic data play a significant role in the model’s final performance and can lead to the improvement of the deep learning systems in musculoskeletal ultrasound. Full article

A conductive meta-atom of toroidal topology is studied both theoretically and experimentally, demonstrating a sharp and highly controllable resonant response. Simulations are performed both for a free-space periodic metasurface and a pair of meta-atoms inserted within a rectangular metallic waveguide. A quasi-dark state with controllable radiative coupling is supported, allowing to tune the linewidth (quality factor) and lineshape of the supported resonance via the appropriate geometric parameters. By conducting a rigorous multipole analysis, we find that despite the strong toroidal dipole moment, it is the residual electric dipole moment that dictates the electromagnetic response. Subsequently, the structure is fabricated with 3D printing and coated with silver paste. Importantly, the structure is planar, consists of a single metallization layer and does not require a substrate when neighboring meta-atoms are touching, resulting in a practical, thin and potentially low-loss system. Measurements are performed in the 5 GHz regime with a vector network analyzer and a good agreement with simulations is demonstrated. Full article

Marine microalgae are considered an important feedstock of multiple valuable metabolic compounds of high biotechnological potential. In this work, the marine microalga Tetraselmis striata was cultivated in different scaled photobioreactors (PBRs). Initially, experiments were performed using two different growth substrates (a modified F/2 and the commercial fertilizer Nutri-Leaf (30% TN—10% P—10% K)) to identify the most efficient and low-cost growth medium. These experiments took place in 4 L glass aquariums at the laboratory scale and in a 9 L vertical tubular pilot column. Enhanced biomass productivities (up to 83.2 mg L⁻¹ d⁻¹) and improved biomass composition (up to 41.8% d.w. proteins, 18.7% d.w. carbohydrates, 25.7% d.w. lipids and 4.2% d.w. total chlorophylls) were found when the fertilizer was used. Pilot-scale experiments were then performed using Nutri-Leaf as a growth medium in different PBRs: (a) a paddle wheel, open, raceway pond of 40 L, and (b) a disposable polyethylene (plastic) bag of 280 L working volume. Biomass growth and composition were also monitored at the pilot scale, showing that high-quality biomass can be produced, with important lipids (up to 27.6% d.w.), protein (up to 45.3% d.w.), carbohydrate (up to 15.5% d.w.) and pigment contents (up to 4.2% d.w. total chlorophylls), and high percentages of eicosapentaenoic acid (EPA). The research revealed that the strain successfully escalated in larger volumes and the biochemical composition of its biomass presents high commercial interest and could potentially be used as a feed ingredient. Full article

Automatically measuring a muscle’s cross-sectional area is an important application in clinical practice that has been studied extensively in recent years for its ability to assess muscle architecture. Additionally, an adequately segmented cross-sectional area can be used to estimate the echogenicity of the muscle, another valuable parameter correlated with muscle quality. This study assesses state-of-the-art convolutional neural networks and vision transformers for automating this task in a new, large, and diverse database. This database consists of 2005 transverse ultrasound images from four informative muscles for neuromuscular disorders, recorded from 210 subjects of different ages, pathological conditions, and sexes. Regarding the reported results, all of the evaluated deep learning models have achieved near-to-human-level performance. In particular, the manual vs. the automatic measurements of the cross-sectional area exhibit an average discrepancy of less than 38.15 mm², a significant result demonstrating the feasibility of automating this task. Moreover, the difference in muscle echogenicity estimated from these two readings is only 0.88, another indicator of the proposed method’s success. Furthermore, Bland–Altman analysis of the measurements exhibits no systematic errors since most differences fall between the 95% limits of agreements and the two readings have a 0.97 Pearson’s correlation coefficient (p < 0.001, validation set) with ICC (2, 1) surpassing 0.97, showing the reliability of this approach. Finally, as a supplementary analysis, the texture of the muscle’s visible cross-sectional area was examined using deep learning to investigate whether a classification between healthy subjects and patients with pathological conditions solely from the muscle texture is possible. Our preliminary results indicate that such a task is feasible, but further and more extensive studies are required for more conclusive results. Full article

The marine microalgae Tetraselmis striata was cultivated in drilling waters with different salinities. Growth substrate optimization was performed while the effects of different pH, temperature, photoperiod and CO₂ flow rate on biomass productivity and its composition were studied. Results showed that the strain grew better in 2.8% drilling waters employing the fertilizer Nutri-Leaf together with ΝaHCO₃. A pH value of 8 resulted in high biomass productivity (79.8 mg L⁻¹ d⁻¹) and biomass composition (proteins 51.2% d.w., carbohydrates 14.6% d.w., lipids 27.8% d.w. and total chlorophylls 5.1% d.w.). The optimum cultivation temperature was found to be 25 ± 1 °C which further enhanced biomass productivity (93.7 mg L⁻¹ d⁻¹) and composition (proteins 38.7% d.w., carbohydrates 20.4% d.w., lipids 30.2% d.w., total chlorophylls 5.1% d.w.). Photoperiod experiments showed that continuous illumination was essential for biomass production. A 10 mL min⁻¹ flow rate of CO₂ lead to biomass productivity of 87.5 mg L⁻¹ d⁻¹ and high intracellular content (proteins 44.6% d.w., carbohydrates 10.3% d.w., lipids 27.3% d.w., total chlorophylls 5.2% d.w.). Applying the optimum growth conditions, the produced biomass presented high protein content with adequate amino acids and high percentages of eicosapentaenoic acid (EPA), indicating its suitability for incorporation into conventional fish feeds. In addition, this study analyzed how functional parameters may influence the uptake of nutrients by Tetraselmis. Full article

Despite the success of hand-crafted features in computer visioning for many years, nowadays, this has been replaced by end-to-end learnable features that are extracted from deep convolutional neural networks (CNNs). Whilst CNNs can learn robust features directly from image pixels, they require large amounts of samples and extreme augmentations. On the contrary, hand-crafted features, like SIFT, exhibit several interesting properties as they can provide local rotation invariance. In this work, a novel scheme combining the strengths of SIFT descriptors with CNNs, namely SIFT-CNN, is presented. Given a single-channel image, one SIFT descriptor is computed for every pixel, and thus, every pixel is represented as an M-dimensional histogram, which ultimately results in an M-channel image. Thus, the SIFT image is generated from the SIFT descriptors for all the pixels in a single-channel image, while at the same time, the original spatial size is preserved. Next, a CNN is trained to utilize these M-channel images as inputs by operating directly on the multiscale SIFT images with the regular convolution processes. Since these images incorporate spatial relations between the histograms of the SIFT descriptors, the CNN is guided to learn features from local gradient information of images that otherwise can be neglected. In this manner, the SIFT-CNN implicitly acquires a local rotation invariance property, which is desired for problems where local areas within the image can be rotated without affecting the overall classification result of the respective image. Some of these problems refer to indirect immunofluorescence (IIF) cell image classification, ground-based all-sky image-cloud classification and human lip-reading classification. The results for the popular datasets related to the three different aforementioned problems indicate that the proposed SIFT-CNN can improve the performance and surpasses the corresponding CNNs trained directly on pixel values in various challenging tasks due to its robustness in local rotations. Our findings highlight the importance of the input image representation in the overall efficiency of a data-driven system. Full article