Browse Articles

Robust multiclass classification of crop leaf diseases using hybrid deep learning and Grad-CAM interpretability

Abstract The key objective of this study is to propose an effective and accurate deep learning (DL) framework to detect and classify diseases in banana, cherry, and tomato leaves. The performance of multiple pre-trained models is compared against a newly presented model.The experiments used a publicly released dataset of healthy and unhealthy leaves from banana, cherry, and tomato plants. This dataset was uniformly split into training, validation, and test sets to obtain consistent and unbiased model evaluations. The data pre-processing also involved pre-processing steps suitable for DL architectures to keep the input the same among all the models.We use several state-of-the-art pre-trained ConvNets models for the baselines, such as EfficientNetV2, ConvNeXt, Swin Transformer, and Vi-Transformer (ViT), to have an outlook on the performance. A new ConvNet-ViT hybrid model combines the ConvNet and ViT layers for local feature extraction and maintaining the global context. The classifier’s performance was reinforced by a 5-fold cross-validation mechanism to avoid overfitting.The proposed Hybrid ConvNet-ViT model outperformed all the compared models evaluated, achieving a testing classification accuracy of 99.29%, which outperforms all the pre-trained models. This finding shows that combining ConvNets’ local feature learning with the capability of global representation of the ViT is effective.The result shows that the Hybrid ConvNet-ViT model is an effective and accurate solution in detecting and classifying plant leaf diseases. Its outstanding performance of the state-of-the-art pre-trained top models positions itself as a solid model for practical agricultural use. Fusing the ConvNet and transformer frameworks jointly is beneficial for improving classification performance in image-based disease detection work.

Machine learning analysis of pharmaceutical cocrystals solubility parameters in enhancing the drug properties for advanced pharmaceutical manufacturing

Psychological and physiological differences related to supportive living situations amongst individuals with physical disabilities

Knowledge of idiopathic normal pressure hydrocephalus among healthcare professionals

Correction: Dual α-amylase and α-glucosidase inhibition by 1,2,4-triazole derivatives for diabetes treatment

A comparative analysis of imaging-based algorithms for detecting focal cortical dysplasia type II in children

Abstract Focal cortical dysplasia (FCD) is the leading cause of drug-resistant epilepsy (DRE) in pediatric patients. Accurate detection of FCDs is crucial for successful surgical outcomes, yet remains challenging due to frequently subtle MRI findings, especially in children, whose brain morphology undergoes significant developmental changes. Automated detection algorithms have the potential to improve diagnostic precision, particularly in cases, where standard visual assessment fails. This study aimed to evaluate the performance of automated algorithms in detecting FCD type II in pediatric patients and to examine the impact of adult versus pediatric templates on detection accuracy. MRI data from 23 surgical pediatric patients with histologically confirmed FCD type II were retrospectively analyzed. Three imaging-based detection algorithms were applied to T1-weighted images, each targeting key structural features: cortical thickness, gray matter intensity (extension), and gray–white matter junction blurring. Their performance was assessed using adult and pediatric healthy controls templates, with validation against both predictive radiological ROIs (PRR) and post-resection cavities (PRC). The junction algorithm achieved the highest median dice score (0.028, IQR 0.038, p < 0.01 when compared with other algorithms) and detected relevant clusters even in MRI-negative cases. The adult template (median dice score 0.013, IQR 0.027) significantly outperformed the pediatric template (0.0032, IQR 0.023) (p < 0.001), highlighting the importance of template consistency. Despite superior performance of the adult template, its use in pediatric populations may introduce bias, as it does not account for age-specific morphological features such as cortical maturation and incomplete myelination. Automated algorithms, especially those targeting junction blurring, enhance FCD detection in pediatric populations. These algorithms may serve as valuable decision-support tools, particularly in settings where neuroradiological expertise is limited.

Psychometric evaluation of the Slovak adaptation of the psychological immune competence inventory (PICI)

Muscle thickness from amplitude mode ultrasound and clinical outcomes in patients with cancer

The use of reduced graphene oxide (rGO) as de-icing agent in engineered cementitious composites (ECC)

Abstract In this study, it is aimed to prevent traffic accidents caused by icing by producing an innovative concrete pavement that is both heatable and has high ductility in order to solve the ductility and icing problems encountered in rigid pavements used in highways. For this aim, rGO-ECCs were produced by adding different proportions of Reduced Graphene Oxide (rGO) to Engineered Cementitious Composites (ECC) known for their high ductility properties. The mechanical, electrical and heating properties of rGO-ECCs in different environments were investigated. Moreover, their energy efficiencies during deicing procedure were compared. Compression test was conducted to all rGO-ECC mixtures. In addition, two-pole conductivity test was performed to measure the resistivity values of rGO-ECCs. Furthermore, heating processes were carried out by applying a carbon-based conductive liquid coating to the surface of rGO-ECCs. Heating processes were applied at room conditions (22C°) and in a cold environment test room (-33C°). The energy efficiency was calculated by measuring the temperature changes of the samples with a thermal camera. TGA-DTA, FT-IR, XRD and SEM–EDX analyses were performed to determine the microstructural properties of rGO-ECCs. Compared to the control (0.0% rGO-ECC) sample, the compressive strength of the 0.6% rGO-ECC sample decreased by approximately 26% and was determined as 52 MPa. The resistivity value of the control sample, which was 4115KΩ.cm, decreased to 49KΩ.cm with the addition of 0.6% rGO-ECC and the conductivity increased. The energy efficiency for temperature change in room (22C°) and cold ambient (-33C°) conditions was calculated as 59.7% and 86.1%, respectively, and the energy efficiency for melting 6 cm of ice in a cold environment (-33C°) was calculated as 63.1%. Finally, a cost analysis was made for ECCs. As a result, it was concluded that rGO can be used as an effective de-icing agent in ECCs.

Melanopsin-mediated image statistics from natural and human-made environments

Abstract Melanopsin-expressing intrinsically photosensitive retinal ganglion cells (ipRGCs) play a critical role in regulating physiological and behavioral responses to light. However, little is known about how melanopsin and ipRGC signals are shaped by the statistical properties of real-world environments. Here, we analyzed the statistics of melanopsin, ipRGC codification of extrinsic and intrinsic photoresponses, and luminance using hyperspectral images of natural and human-made scenes under daylight illumination. The statistics were obtained by modeling human retinal receptive fields from current knowledge about ipRGC anatomy and physiology. Our findings reveal that human-made environments exhibit significantly higher melanopsin, luminance, and ipRGC excitations compared to natural environments. This difference is linked to higher reflectance values in human-made environments. In natural scenes, luminance contrasts were higher than melanopsin and ipRGC contrasts across most of the range. Melanopsin contrast was largely independent of excitation and was significantly reduced for larger receptive fields. Differences between ipRGC codification models suggest an interaction between input weighting and environmental structure. These results indicate that modifications of natural regularities by human-made environments could affect ipRGC-driven physiology in everyday life and may deviate from the evolutionary constraints that shaped ipRGC function.

Tuning of task-relevant stiffness in multiple directions

Abstract In contrast to robots, humans can rapidly and elegantly modulate the impedance of their arms and hands during initial contact with objects. Anticipating collisions by setting mechanical impedance to counter near-instantaneous changes in force and displacement is one reason we excel at manipulating objects. We investigated the ability to set impedance in an object interaction task with rapid changes in force and displacement, like those encountered during manipulation in different directions. Subjects ( n  = 20) predictively co-activated antagonist muscles to adjust one component of the impedance – stiffness – to match the task demands before the movement began, irrespective of movement direction. Subjects adopted the minimal stiffness needed to complete the task, but when pushed to the most difficult condition, they were limited by their ability to produce high stiffness rather than large force. This robust and simple strategy ensured task success at the expense of energy efficiency. Our results confirm the ability of humans to predictively set and control mechanical impedance in task-relevant directions in anticipation of breaking contact. This offers the prospect that future investigations will find neural correlates of impedance, which in turn, could improve the ability of neuro-prosthetic limbs to interact with objects.

Numerical and experimental investigation of the solar air heater with latent heat storage and fin

Author Correction: Changes in human mandibular shape during the Terminal Pleistocene-Holocene Levant

Depressive symptoms and associated factors among patients with diabetes in public primary healthcare facilities in Kuwait city, 2024

The critical dimension of memory engrams and an optimal number of senses

Abstract In this work, we analyse the fundamental question: how many senses are optimal for memory and learning. To answer this question, we introduce and analyse a novel kinetic model of memory engrams. The model, built on basic general principles and phenomenology, captures the engrams’ emergence and evolution driven by their interaction with external environment, learning, and forgetting. We derive the corresponding kinetic equation governing the dynamics and evolution of engrams over time. We then solve this equation analytically and numerically through Monte Carlo simulations. We observe the formation of a steady state with a steady number of different engrams covering a fraction of the conceptual space. We analyze the impact of the dimension of the conceptual space on the steady state and discover the existence of a critical dimension, at which the number of different engrams is maximal; we provide a theoretical explanation of this observation. If each feature is associated with a different sense, the critical dimension corresponds to an optimal number of senses for a system aiming at keeping the maximal number of different concepts in its memory. We also reveal an apparent tension between the system’s receptivity to new stimuli and concept sharpness—the higher the receptivity, the less sharp the learned concept becomes.

Exploring organic compound preservation through long-term in situ experiments in the Atacama desert and the relevance for Mars

Abstract The preservation of organic compounds under extreme environmental conditions remains a critical challenge for both terrestrial ecology applications on Earth and astrobiology. In a novel long-term field experiment over 8 months, we exposed biomolecules and a model organism to natural hyperarid conditions of the Atacama Desert, one of the best Mars analog environments. We used custom-designed sample plates for long-term exposure to simulate environmental stresses that biomolecules are exposed naturally in a hyperarid environment. The multiple stressors included extreme temperature fluctuations, associated humidity changes, and intense solar irradiation. Our field experiment complements and extends the insights obtained from previously conducted short-term laboratory experiments. To investigate biomolecule stability, we embedded adenosine triphosphate (ATP), chlorophyll-a, and the cyanobacterium Chrooccoccidiopsis in various Mars-relevant sediments with addition of chloride and perchlorate. Our findings, which include the rapid degradation of these biomolecules, the detection of more stable degradation products, and the identification of non-enzymatic degradation pathways, reveal the critical influence of substrate and salt types on biomolecule stability. Valuable insights into biosignature preservation under extreme terrestrial conditions and a better understanding of organic signal interpretations were gained, which will provide critical insights for future Mars missions, especially when searching for past or present life.

Variabilities of ground motion in 2023 Mw 7.8 Pazarcık earthquake in Turkey and 2008 Mw 7.9 Wenchuan earthquake in China

Non-contact ultrasound to assist laser additive manufacturing

Temporal relationship between dancer’s body movements and music beats in classical ballet

Anisotropy reveals contact sliding and aging as a cause of post-seismic velocity changes

Abstract Rocks exhibit astonishing time-dependent mechanical properties, like memory of experienced stress or slow dynamics, a transient recovery of stiffness after a softening induced by almost any type of loading. This softening and transient recovery is observed in the subsurface and in buildings after earthquakes, or in laboratory samples. Here, we investigate the anisotropy of nonlinear elastic effects in a sandstone sample under uniaxial loading. We report that slow dynamics is observed independently of propagation direction, while the acoustoelastic effect shows the expected anisotropy originating from the opening and closing of cracks. From this, we argue that slow dynamics is caused by the sliding of oblique grain-to-grain contacts and the resulting changes in frictional properties, as empirically described by rate-and-state friction and observed in laboratory experiments across block contacts. We establish a connection between the nonclassical nonlinearity of heterogeneous materials and the framework of rate-and-state friction, providing an explanation for the elusive origin of slow dynamics, and adding a different perspective for monitoring very early stages of material failure when deformation is still distributed in the bulk and begins to coalesce towards a fracture.