Search Results (9,750)

Unmanned aerial vehicles (UAVs) have become a tool for solving numerous practical tasks. UAV sensors provide images and videos for on-line or off-line data processing for object localization, classification, and tracking due to the use of trained convolutional neural networks (CNNs) and artificial intelligence. However, quality of images acquired by UAV-based sensors is not always perfect due to many factors. One of them could be noise arising because of several reasons. Its presence, especially if noise is intensive, can make significantly worse the performance characteristics of CNN-based techniques of object localization and classification. We analyze such degradation for a set of eleven modern CNNs for additive white Gaussian noise model and study when (for what noise intensity and for what CNN) the performance reduction becomes essential and, thus, special means to improve it become desired. Representatives of two most popular families, namely the block matching 3-dimensional (BM3D) filter and DRUNet denoiser, are employed to enhance images under condition of a priori known noise properties. It is shown that, due to preliminary denoising, the CNN performance characteristics can be significantly improved up to almost the same level as for the noise-free images without CNN retraining. Performance is analyzed using several criteria typical for image denoising, object localization and classification. Examples of object localization and classification are presented demonstrating possible object missing due to noise. Computational efficiency is also taken into account. Using a large set of test data, it is demonstrated that: (1) the best results are usually provided for SSD Mobilenet V2 and VGG16 networks; (2) the performance characteristics for cases of applying BM3D filter and DRUNet denoiser are similar but the use of DRUNet is preferable since it provides slightly better results. Full article

Antiviral peptides (AVPs) are promising therapeutic candidates, yet the rapid growth of sequence data and the field’s emphasis on predictors have left a gap: the lack of an integrated view linking peptide chemistry with biological context. Here, we map the AVP landscape through interactive data mining using Half-Space Proximal Networks (HSPNs) and Metadata Networks (MNs) in the StarPep toolbox. HSPNs minimize edges and avoid fixed thresholds, reducing computational cost while enabling high-resolution analysis. A threshold-free HSPN resolved eight chemically and biologically distinct communities, while MNs contextualized AVPs by source, function, and target, revealing structural–functional relationships. To capture diversity compactly, we applied centrality-guided scaffold extraction with redundancy removal (90–50% identity), producing four representative subsets suitable for modeling and similarity searches. Alignment-free motif discovery yielded 33 validated motifs, including 10 overlapping with reported AVP signatures and 23 apparently novel. Motifs displayed category-specific enrichment across antimicrobial classes, and sequences carrying multiple motifs (≥4–5) consistently showed higher predicted antiviral probabilities. Beyond computational insights, scaffolds provide representative “entry points” into AVP chemical space, while motifs serve as modular building blocks for rational design. Together, these resources provide an integrated framework that may inform AVP discovery and support scaffold- and motif-guided therapeutic design. Full article

Molecular networking (MN) is a novel mass spectrometry data analysis method that has advanced significantly in recent years and has rapidly emerged as a popular technique. By visualizing the connections between structurally similar compounds in mass spectra, MN greatly enhances the efficiency with which harmful substances and bioactive ingredients in cosmetics are screened. In this review, we summarize the principles and main categories of MN technology and systematically synthesize its progress in cosmetic testing applications based on 83 recent studies (2020 to 2025). These applications include screening banned additives, analyzing complex matrix components, and identifying efficacy-related ingredients. We highlight MN’s successful application in detecting prohibited substances, such as synthetic dyes and adulterants, with limits of detection (LOD) as low as 0.1–1 ng/g, even in complex matrices, such as emulsions and colored products. MN-guided isolation has enabled the structural elucidation of over 40 known and novel compounds in the analysis of natural ingredients. We also discuss current challenges, such as limitations in instrument sensitivity, matrix effects, and the lack of cosmetic-specific component databases. Additionally, we outline future prospects for expanding MN’s application scope in cosmetic testing and developing it toward computer-aided intelligence. This review aims to provide valuable references for promoting innovation in cosmetic testing methods and strengthening quality control in the industry. Full article

In this article, we present a novel reinforcement learning-based framework in the discrete cosine transform to achieve better image steganography. First, the input image is divided into several blocks to extract semantic and structural features, evaluating their suitability for data embedding. Second, the Proximal Policy Optimization algorithm (PPO) is introduced in the block selection process to learn adaptive embedding policies, which effectively balances image fidelity and steganographic security. Moreover, the Deep Q-network (DQN) is used for adaptively adjusting the weights of the peak signal-to-noise ratio, structural similarity index, and detection accuracy in the reward formulation. Experimental results on the BOSSBase dataset confirm the superiority of our framework, achieving both lower detection rates and higher visual quality across a range of embedding payloads, particularly under low-bpp conditions. Full article

The Nile tilapia (Oreochromis niloticus), a key aquaculture species, displays marked sexual growth dimorphism, with males growing faster than females. This process is governed by intricate interactions between antagonistic regulators, including transcription factors, growth factors, and steroid hormones, operating through sex-specific developmental pathways. While circular RNAs (circRNAs) are known to modulate gene expression by sponging microRNAs (miRNAs), their role in teleost sex differentiation remains poorly understood. To address this gap, we profiled circRNA expression in tilapia gonads by constructing six circRNA libraries from testes and ovaries of 180 days after hatching (dah) fish, followed by high-throughput sequencing. We identified 6564 gonadal circRNAs distributed across all 22 linkage groups, including 226 differentially expressed circRNAs (DECs; 108 testis-biased, 118 ovary-biased). Functional enrichment analysis linked their host genes to critical pathways such as cAMP signaling, cell adhesion molecules, and—notably—sexual differentiation processes (e.g., estrogen signaling, oocyte meiosis, and steroid hormone biosynthesis). Furthermore, we deciphered competing endogenous RNA (ceRNA) networks, uncovering circRNA–miRNA–mRNA interactions targeting germ cell determinants, sex-specific transcription factors, and steroidogenic enzymes. This study provides the first systematic exploration of circRNA involvement in tilapia sex differentiation and gonadal differentiation, offering novel insights into the post-transcriptional regulation of sexual dimorphism. Our findings advance the understanding of circRNA biology in fish and establish a framework for future studies on aquaculture species with similar reproductive strategies. Full article

The proliferation of social network data has unlocked unprecedented opportunities for extensive, data-driven exploration of human behavior. The structural intricacies of social networks offer insights into various computational social science issues, particularly concerning social influence and information diffusion. However, modeling large-scale social network data comes with computational challenges. Though large language models make it easier than ever to model textual content, any advanced network representation method struggles with scalability and efficient deployment to out-of-sample users. In response, we introduce a novel approach tailored for modeling social network data in user-detection tasks. This innovative method integrates localized social network interactions with the capabilities of large language models. Operating under the premise of social network homophily, which posits that socially connected users share similarities, our approach is designed with scalability and inductive capabilities in mind, avoiding the need for full-graph training. We conduct a thorough evaluation of our method across seven real-world social network datasets, spanning a diverse range of topics and detection tasks, showcasing its applicability to advance research in computational social science. Full article

Neurodevelopmental disorders (NDDs), including autism spectrum disorder, intellectual disability, and attention-deficit/hyperactivity disorder, are genetically and phenotypically heterogeneous conditions affecting millions worldwide. High-throughput omics technologies—transcriptomics, proteomics, metabolomics, and epigenomics—offer a unique opportunity to link genetic variation to molecular and cellular mechanisms underlying these disorders. However, the high dimensionality, sparsity, batch effects, and complex covariance structures of omics data present significant statistical challenges, requiring robust normalization, batch correction, imputation, dimensionality reduction, and multivariate modeling approaches. This review provides a comprehensive overview of statistical frameworks for analyzing high-dimensional omics datasets in NDDs, including univariate and multivariate models, penalized regression, sparse canonical correlation analysis, partial least squares, and integrative multi-omics methods such as DIABLO, similarity network fusion, and MOFA. We illustrate how these approaches have revealed convergent molecular signatures—synaptic, mitochondrial, and immune dysregulation—across transcriptomic, proteomic, and metabolomic layers in human cohorts and experimental models. Finally, we discuss emerging strategies, including single-cell and spatially resolved omics, machine learning-driven integration, and longitudinal multi-modal analyses, highlighting their potential to translate complex molecular patterns into mechanistic insights, biomarkers, and therapeutic targets. Integrative multi-omics analyses, grounded in rigorous statistical methodology, are poised to advance mechanistic understanding and precision medicine in NDDs. Full article

Rapid urbanization and frequent extreme events have made urban flooding a growing threat to residents. This issue is acute in old urban districts, where extremely limited land resources, outdated standards and poor infrastructure have led to inadequate drainage and uneven pipe settlement, heightening flood risk. This study applies InfoWorks ICM Ultimate (version 21.0.284) to simulate flooding in a typical old urban district for six return periods. A risk assessment was carried out, flood causes were analyzed, and mitigation strategies were evaluated to reduce inundation and cost. Results show that all combined schemes outperform single-measure solutions. Among them, the green roof combined with pipe optimization scheme eliminated high-risk and medium-risk areas, while reducing low-risk areas by over 78.23%. It also lowered the ponding depth at key waterlogging points by 70%, significantly improving the flood risk profile. The permeable pavement combined with pipe optimization scheme achieved similar results, reducing low-risk areas by 77.42% and completely eliminating ponding at key locations, although at a 50.8% higher cost. This study underscores the unique contribution of cost-considered gray-green infrastructure retrofitting in old urban areas characterized by land scarcity and aging pipeline networks. It provides a quantitative basis and optimization strategies for refined modeling and multi-strategy management of urban waterlogging in such regions, offering valuable references for other cities facing similar challenges. The findings hold significant implications for urban flood control planning and hydrological research, serving as an important resource for urban planners engaged in flood risk management and researchers in urban hydrology and stormwater management. Full article

Accurate and reliable vehicle localization is essential for autonomous driving in complex outdoor environments. Traditional feature-based visual–inertial odometry (VIO) suffers from sparse features and sensitivity to illumination, limiting robustness in outdoor scenes. Deep learning-based optical flow offers dense and illumination-robust motion cues. However, existing methods rely on simple bidirectional consistency checks that yield unreliable flow in low-texture or ambiguous regions. Global navigation satellite system (GNSS) measurements can complement VIO, but often degrade in urban areas due to multipath interference. This paper proposes a multi-sensor fusion system that integrates monocular VIO with GNSS measurements to achieve robust and drift-free localization. The proposed approach employs a hybrid VIO framework that utilizes a deep learning-based optical flow network, with an enhanced consistency constraint that incorporates local structure and motion coherence to extract robust flow measurements. The extracted optical flow serves as visual measurements, which are then fused with inertial measurements to improve localization accuracy. GNSS updates further enhance global localization stability by mitigating long-term drift. The proposed method is evaluated on the publicly available KITTI dataset. Extensive experiments demonstrate its superior localization performance compared to previous similar methods. The results show that the filter-based multi-sensor fusion framework with optical flow refined by the enhanced consistency constraint ensures accurate and reliable localization in large-scale outdoor environments. Full article

Background/Objectives: Despite their ecological significance, DNA barcoding data for African rainforest Orthoptera remain underrepresented globally, limiting progress in species discovery, biodiversity assessment, and conservation. This study aimed to generate molecular data for morphologically identified Serpusia Karsch, 1891 species to evaluate their taxonomic status. Methods: Specimens were collected from multiple sites in Cameroon and analyzed using DNA barcoding with COI-5P and 16S rDNA markers. Species delimitation was performed with Automatic Barcode Gap Discovery, and phylogenetic relationships were inferred using Maximum Likelihood and Bayesian Inference. Additionally, external morphology and the male phallic complex were examined. Results: Molecular analyses delineated 19 MOTUs, five corresponding to Serpusia opacula, seven to Serpusia succursor and the remainder to outgroups. Similarity-based assignments matched these MOTUs to 19 BINs. Phylogenetic reconstruction revealed S. opacula and S. succursor as two genetically distinct clades, with the S. opacula group more closely related to Aresceutica Karsch, 1896 than to the S. succursor group. Accordingly, we established a new genus, Paraserpusia gen. nov., to accommodate S. succursor. Within the S. opacula group, five species are recognized: one previously described (S. opacula) and four new species (S. kennei sp. nov., S. missoupi sp. nov., S. seinoi sp. nov., and S. verhaaghi sp. nov.). The former S. succursor, now Paraserpusia succursor, is divided into six well-supported lineages, five of which are formally described here (P. hoeferi sp. nov., P. husemanni sp. nov., P. kekeunoui sp. nov., P. tamessei sp. nov., and P. tindoi sp. nov.). A haplotype network based on COI-5P sequences corroborates three major clades corresponding to the S. opacula group, the S. succursor group, and Aresceutica. Diagnostic morphological differences between Serpusia and Paraserpusia are consistently supported across characters. Conclusions: This integrative approach reveals substantial hidden diversity within Serpusia and highlights the importance of combining molecular and morphological data to uncover and formally describe previously overlooked taxa. Full article

In field exploration, disaster rescue, and complex terrain operations, the accuracy of ground type recognition directly affects the walking stability and task execution efficiency of legged robots. To address the problem of terrain recognition in complex ground environments, this paper proposes a high-precision classification method based on single-leg triaxial force signals. The method first employs a one-dimensional convolutional neural network (1D-CNN) module to extract local temporal features, then introduces a long short-term memory (LSTM) network to model long-term and short-term dependencies during ground contact, and incorporates a convolutional block attention module (CBAM) to adaptively enhance the feature responses of critical channels and time steps, thereby improving discriminative capability. In addition, an improved whale optimization algorithm (iBWOA) is adopted to automatically perform global search and optimization of key hyperparameters, including the number of convolution kernels, the number of LSTM units, and the dropout rate, to achieve the optimal training configuration. Experimental results demonstrate that the proposed method achieves excellent classification performance on five typical ground types—grass, cement, gravel, soil, and sand—under varying slope and force conditions, with an overall classification accuracy of 96.94%. Notably, it maintains high recognition accuracy even between ground types with similar contact mechanical properties, such as soil vs. grass and gravel vs. sand. This study provides a reliable perception foundation and technical support for terrain-adaptive control and motion strategy optimization of legged robots in real-world environments. Full article

In recent years, agriculture has increasingly faced shocks related to climate change, pathogen outbreaks, and geopolitical instability, highlighting the need for sustainable regeneration strategies. This study develops an innovative Multi-Criteria Decision-Making (MCDM) framework that integrates the Delphi method, the Analytic Network Process (ANP), and the Aggregated Decision-Making (ADAM) method—the first application of this combination in the context of agricultural regeneration. The framework was applied to the Apulia region (Italy), heavily affected by the Xylella fastidiosa epidemic, and evaluated alternative crops across 30 economic, environmental, and socio-cultural sub-criteria. Results indicate that carob, walnut, and pistachio outperform other options by combining strong economic viability, climate resilience, and cultural compatibility. To mitigate the risks of monoculture, crop diversification strategies based on high-ranked alternatives are recommended. Sensitivity analysis confirmed the robustness of results, and the framework demonstrates high scalability, offering a transparent tool for policymakers in regions facing similar agricultural crises. Full article

This paper presents a novel adversarial Convolutional Neural Network (ConvNet) method for generating adversarial perturbations in 3D point clouds, enabling gradient-free robustness evaluation of object detection systems at inference time. Unlike existing iterative gradient methods, our approach embeds the ConvNet directly into the detection pipeline at the voxel feature level. The ConvNet is trained to maximize detection loss while maintaining perturbations within sensor error bounds through multi-component loss constraints (intensity, bias, and imbalance terms). Evaluation on a Sparsely Embedded Convolutional Detection (SECOND) detector with the KITTI dataset shows 8% overall mean Average Precision (mAP) degradation, while CenterPoint on NuScenes exhibits 24% weighted mAP reduction across 10 object classes. Analysis reveals an inverse relationship between object size and adversarial vulnerability: smaller objects (pedestrians: 13%, cyclists: 14%) show higher vulnerability compared to larger vehicles (cars: 0.2%) on KITTI, with similar patterns on NuScenes, where barriers (68%) and pedestrians (32%) are most affected. Despite perturbations remaining within typical sensor error margins (mean L2 norm of 0.09% for KITTI, 0.05% for NuScenes, corresponding to 0.9–2.6 cm at typical urban distances), substantial detection failures occur. The key novelty is training a ConvNet to learn effective adversarial perturbations during a one-time training phase and then using the trained network for gradient-free robustness evaluation during inference, requiring only a forward pass through the ConvNet (1.2–2.0 ms overhead) instead of iterative gradient computation, making continuous vulnerability monitoring practical for autonomous driving safety assessment. Full article

Color grading of boiled shrimp is a critical factor influencing market price, yet the process is usually conducted visually by buyers such as middlemen and processing plants. This subjective practice raises concerns about accuracy, impartiality, and fairness, often resulting in disputes with farmers. To address this issue, this study proposes a standardized and automated grading approach based on image processing and artificial intelligence. The method requires only a photograph of boiled shrimp placed alongside a color grading ruler. The grading process involves two stages: segmentation of shrimp and ruler regions in the image, followed by color comparison. For segmentation, deep learning models based on Mask R-CNN with a Feature Pyramid Network backbone were employed. Four model configurations were tested, using ResNet and ResNeXt backbones with and without a Boundary Loss function. Results show that the ResNet + Boundary Loss model achieved the highest segmentation performance, with IoU scores of 91.2% for shrimp and 87.8% for the color ruler. In the grading step, color similarity was evaluated in the CIELAB color space by computing Euclidean distances in the L (lightness) and a (red–green) channels, which align closely with human perception of shrimp coloration. The system achieved grading accuracy comparable to human experts, with a mean absolute error of 1.2, demonstrating its potential to provide consistent, objective, and transparent shrimp quality assessment. Full article

The selection of an appropriate technological process is essential to achieve optimal results in manufacturing companies. This affects quality, efficiency and competitiveness. In the modern industry, multi-criteria decision-making (MCDM) methods are increasingly used to evaluate, optimize and solve various manufacturing challenges. In this review article, existing methodologies and patents related to optimization and decision making are investigated. The main characteristics and applications of the methods are outlined. The purpose of this article is to provide a systematic review and evaluation of the main MCDM methods used in industrial practice, including through an analysis of relevant methodologies and patents. The methodology involves a structured literature and patent review, focusing on applications of widely used MCDM techniques such as the AHP (analytic hierarchy process), ANP (analytic network process), FUCOM (full consistency method), TOPSIS (technique for order preference by similarity to ideal solution), and VIKOR (višekriterijumsko kompromisno rangiranje). The analysis outlines each method’s strengths, limitations and areas of applicability. Special attention is given to the potential of the FUCOM for process evaluation in manufacturing. The findings are intended to guide researchers and practitioners in selecting appropriate decision-making tools based on specific industrial contexts and objectives. In conclusion, from the comparative analysis made, the methodologies reveal their advantages and disadvantages as well as limitations that arise in their application. Full article