Search Results (448)

To address the challenges of low detection accuracy and weak generalization in identifying mining-induced ground fissures from UAV imagery, caused by their slender and discontinuous morphology, complex background clutter, and multi-scale surface features, this paper proposes an enhanced deep semantic segmentation model, SAG-DeepLabV3+ (with Spatial Vision Transformer, Attention mechanisms, and Adaptive Gated Fusion). Specifically, to enhance global context modeling and fine boundary delineation, we introduce a Spatial Vision Transformer (SVT) branch within the Atrous Spatial Pyramid Pooling (ASPP) module. We further employ a dual attention mechanism, sequentially combining Squeeze-and-Excitation (SE) and a Convolutional Block Attention Module (CBAM), for progressive channel and spatial feature refinement. Moreover, an Adaptive Gated Fusion (AGF) module is designed to dynamically optimize the fusion of multi-level decoder features. Experiments on a dedicated UAV-based mining fissure dataset comprising 1280 annotated images show that SAG-DeepLabV3+ achieves a state-of-the-art mean Intersection over Union (mIoU) of 79.52% (with Xception backbone) and 79.19% (with lightweight MobileNetV2 backbone), surpassing DeepLabV3+, U-Net, and PSPNet by a significant margin. Furthermore, by leveraging transfer learning (pre-training on the public CrackVision12K dataset and fine-tuning on our mining fissure dataset), the model’s mIoU is further elevated to 82.04%, demonstrating superior generalization capability. The proposed SAG-DeepLabV3+ effectively balances high accuracy with operational efficiency, fulfilling the potential demand for lightweight automated fissure monitoring under resource-limited field deployments, and lays a foundation for subsequent real-time on-site deployment verification. Full article

Low-light color image enhancement remains a challenging task for vision-based decision systems, which must simultaneously address illumination correction, noise suppression, contrast recovery, and color preservation from a single degraded observation. This study proposes ResGASP-GAN, a GAN-based low-light image enhancement framework built around a residual-output generator that integrates batch-size-independent normalization, multi-scale contextual aggregation, and channel-wise feature recalibration within a conditional adversarial setting. Group Normalization is integrated into the proposed generator to reduce dependence on batch statistics during small-batch training, while a Squeeze-and-Excitation (SE) module adaptively enables channel-wise feature recalibration and helps preserve structural and chromatic information. The proposed generator uses reflection-padded convolutions to reduce boundary artifacts, and it features a multi-scale bottleneck composed of dilated residual blocks and Atrous Spatial Pyramid Pooling to capture spatially varying illumination patterns. The model is optimized using a compound objective that combines an adversarial term with an ${\ell }_1$ reconstruction loss, balancing perceptual realism with pixel-level fidelity. Experimental evaluation employed the LOL-v1, LOL-v2-Real, and LOL-v2-Synthetic datasets using reference-based metrics: PSNR, SSIM, and LPIPS. No-reference perceptual metrics were also used, including NIQE and BRISQUE. The results indicate that the proposed method achieves competitive structural similarity and visual image quality on LOL-v2-Real, competitive reconstruction performance on LOL-v1, and good generalization on LOL-v2-Synthetic, with the second-best metrics of PSNR = 22.30 dB, SSIM = 0.9154, and LPIPS = 0.1022 among the reported methods on this last dataset. In contrast, using no-reference metrics, this study achieved very good results, with the lowest BRISQUE of 10.4540 and a competitive NIQE of 4.4396, providing high-quality visual–perceptual reconstruction. Overall, the proposed architecture provides a competitive GAN-based alternative for low-light image enhancement, combining residual connections, multi-scale contextual modeling, and channel-wise feature refinement. This architecture provides the lowest inference time over all discussed models, which is highly required in real-time outdoor applications such as robot navigation and mapping. Full article

The May–September 2021 cryptocurrency-mining prohibition in the People’s Republic of China removed an industry that had become heavily concentrated in Chinese jurisdictions over the years preceding the ban. Using the full population of 97,005 Bitcoin blocks mined between May 2020 and February 2022, retrieved from the mempool.space block API, with pool-operator corporate domicile and China-linkage coded from publicly documented corporate-parent records, we estimate the post-prohibition shift in pool-attributed block share. On the top-13 panel covering approximately ninety-five per cent of attributed blocks, a two-way fixed-effects projection records a differential China-linked share movement of $-5.45$ percentage points per pool after the 24 September 2021 NDRC and PBOC enforcement event (cluster-robust SE 2.55; CRV1 p = 0.054, not significant at five per cent; wild cluster bootstrap p = 0.014 and exact-permutation p = 0.023 reported as small-sample diagnostics, not relied on for significance; under a permutation conditioned on baseline share the latter rises to p = 0.100). This paper’s primary magnitudes are the aggregate compositional movements, which do not depend on the cluster-level test. Aggregate China-linked share falls from 78.47 to 60.01 per cent between the placebo (May–October 2020) and persistence (November 2021–February 2022) windows; non-China-linked share rises from 3.50 to 27.02 per cent, with Foundry USA absorbing approximately 15.7 percentage points (approximately 67 per cent of the non-China-linked share gain). The within-non-China-subset normalised Herfindahl–Hirschman Index falls from 10,000 to 4191, consistent with rising within-subset dispersion among non-Chinese-domiciled operators. Among the identified non-China-linked operators absorbing material share in the analytical panel, the displacement is concentrated at named legal persons, chiefly with documented United States and European corporate domicile; the residual excluded/unknown category is not used to support that claim. Full article

Robust monitoring of weld pool dynamics is critical for automated arc welding; however, single-modality sensors are frequently constrained by severe optical interference and high-frequency environmental noise. To address these limitations, we propose M2WPR-Net, a novel multimodal framework that synergizes visual and acoustic signals for simultaneous weld width regression and physical quality classification. The architecture employs a dual-stream ResNet50 backbone to process heterogeneous sensory data. Specifically, the visual stream utilizes a Convolutional Block Attention Module (CBAM) to suppress intense arc glare and localize the weld pool. Concurrently, the acoustic stream transforms 1D audio sequences into 2D Gramian Angular Summation Field (GASF) textures, which are subsequently refined by Squeeze-and-Excitation (SE) networks to isolate target frequency channels. A central contribution of this study is a bidirectional cross-modal attention mechanism based on Query–Key–Value (Q-K-V) matrix operations. Overcoming the shortcomings of static feature concatenation, this module dynamically aligns the modalities, enabling acoustic cues to guide visual feature extraction and vice versa, thereby mitigating information bottlenecks. Optimized via a joint multi-task loss function, the proposed M2WPR-Net significantly outperforms existing single-modal and conventional fusion baselines. Experimental results demonstrate that the network achieves a Mean Absolute Error (MAE) of 0.18 mm for width prediction and a 93.5% accuracy in penetration state classification, confirming its resilience and practical applicability in complex industrial welding environments. Full article

Organic memristors are attractive building blocks for neuromorphic computing owing to their intrinsic synaptic functionalities and solution-processability. However, their operational instability remains a major challenge, primarily arising from poorly controlled semiconductor crystallization and stochastic conductive filament formation. Here, we report a high-performance solution-processed organic memristor based on a TIPS-pentacene/PMMA/CdSe-ZnS quantum-dot hybrid system, in which a dual-engineering strategy is employed to simultaneously regulate film crystallization and filament dynamics. Specifically, the PMMA matrix templates the molecular ordering of TIPS-pentacene to improve film uniformity and crystallinity, while CdSe/ZnS quantum dots locally modulate the electric field to direct and confine conductive filament formation. As a result, the device exhibits ultralow and highly uniform switching voltages (0.473 V for set and −0.430 V for reset), suppressed device-to-device variation, long retention exceeding 10⁴ s, and endurance over 1200 switching cycles. In addition, the memristor supports multilevel data storage and successfully emulates key synaptic functions, including long-term potentiation/depression, paired-pulse facilitation, and spike-timing-dependent plasticity. This work provides a materials-level strategy for achieving reliable and low-power organic memristors, offering a viable route toward high-density nonvolatile memory and neuromorphic computing hardware. Full article

To address the issue of radioactive waste generated by the large-scale promotion and use of nuclear energy, safety evaluations of disposal sites in various surrounding rocks are essential. These evaluations are a prerequisite for ensuring the long-term safe disposal of radioactive waste. This study focuses on the blocking capacity of clay rocks concerning the advection–dispersion behavior of representative radionuclides such as ¹³⁷Cs, ⁷⁹Se, and ⁹⁹Tc. It further examines the effects of competitive adsorption that arise when these three radionuclides coexist. (Since ⁷⁹Se is difficult to obtain, ⁷⁵Se was used as a substitute nuclide. In the mixed-nuclide experiments, the stable isotope Re was used to replace ⁹⁹Tc.) The experimental findings revealed that competitive adsorption can significantly reduce the adsorption capability of clay rocks for ¹³⁷Cs and ⁷⁹Se, altering the adsorption mechanism. During the advection–dispersion process, the weak adsorption sites of ¹³⁷Cs and ⁷⁹Se on clay rocks become active after the strong adsorption sites are preferentially occupied, resulting in a decline in both adsorption quantity and rate. In the case of ⁹⁹Tc, competitive adsorption weakens the effect of anion repulsion, leading to a reduction in the immobile liquid regions (θ_im). Full article

Reliable condition monitoring of wind turbine gearboxes is critical to reducing unplanned downtime and maintenance costs in wind farms. However, this task presents significant challenges due to the non-stationary nature of vibration signals, in which fault-relevant features are sparsely and unevenly distributed across the time–frequency map. Although time–frequency analysis has been widely adopted to represent nonlinear and non-stationary vibration signals, existing deep learning methods typically process the full spectrogram directly, without distinguishing redundant or uninformative regions. This leads to high input dimensionality and exposes the model to substantial spectral noise. Consequently, it increases computational burden and potentially reduces the diagnostic reliability. To address this issue, this paper proposes a two-stage hybrid framework based on complementary selection mechanisms operating on two distinct feature spaces. In the first stage, eXtreme Gradient Boosting (XGBoost) importance scores are used to identify and permanently prune uninformative time–frequency features from the input spectrogram, reducing the input map size by 25%. In the second stage, a Squeeze-and-Excitation (SE) block, inserted after the deepest residual layer, performs soft channel-wise recalibration of the abstract feature maps produced by the residual convolutional neural network (ResCNN), thereby amplifying discriminative representations prior to classification. The proposed method was evaluated in an eight-class variable-speed fault classification task using the MCC5-THU benchmark, where data were collected from a 2.2 kW motor-driven gearbox test rig. The proposed method achieves a mean accuracy of 97.81% ± 0.33% under 5-fold stratified cross-validation (CV), while reducing classifier training time by approximately 23% compared to a baseline model trained on the full spectrogram. These results demonstrate that explicit input-level spectrogram pruning, combined with model-level channel attention, yields a robust and computationally efficient diagnostic framework for wind turbine gearbox condition monitoring. Full article

Perseverative cognition—frequently expressed through obsessive–compulsive symptoms—is a recognized trans-diagnostic risk factor for psychopathology, with documented links to depression and suicidal ideation. While the literature suggests that altered autonomic arousal in depression may relate to suicidal risk, the conditional architecture linking these variables remains insufficiently explored. This study examined whether suicidal ideation mediates the association between depression and autonomic arousal and whether obsessive–compulsive symptoms moderate the initial pathways within this mechanism. A sample of 120 university students (61.8% female; M_age = 28.6, SD_age = 10.71) completed the Symptom Checklist-90-Revised (SCL-90-R), from which the depression, obsessive–compulsive, and suicidal ideation scores were derived; notably, the suicidal ideation score was calculated from specific items (15 and 59) embedded within the depression subscale. All participants underwent a psychophysiological evaluation to record baseline Electrodermal Activity (EDA) as an index of tonic autonomic arousal. Results indicated that depression significantly predicted suicidal ideation (B = 0.05, SE = 0.006, p < 0.001, 95% CI [0.03, 0.06]), which, in turn, was a significant predictor of autonomic arousal (B = 0.17, SE = 0.08, p = 0.03, 95% CI [0.02, 0.33]). Additionally, the mediation analysis revealed that depression showed an indirect statistical association with autonomic arousal through suicidal ideation (Bootstrapped 95% CI [0.01, 0.12]). Concurrently, obsessive–compulsive symptoms significantly moderated the psychological link between depression and suicidal ideation (B = 0.20, SE = 0.04, p < 0.001, 95% CI [0.04, 0.37]), accounting for 38.40% of the variance (F (4, 115) = 22.17, p < 0.001). Sensitivity analyses, conducted by re-calculating the models after excluding items 15 and 59 from the depression score, confirmed the robustness of these findings against potential psychometric overlap. However, the formal Index of Moderated Mediation was not statistically significant (0.0000 (95% Boot CI [−0.0001, 0.0001]), explicitly demonstrating that this psychological moderation did not translate into a significant conditional indirect effect on peripheral autonomic arousal. These findings suggest that while obsessive cognitive patterns significantly intensify the psychological burden of depression into thoughts of death, their concurrent association with peripheral physiological parameters remains strictly exploratory in non-clinical cohorts, underscoring a nuanced underscoring a nuanced ‘blocking’ effect—both emotional and behavioral—where obsessive rigidity functions as a mechanism of paralysis that, while generating profound subjective distress, fails to translate into a corresponding reactive physiological response of the autonomic nervous system. Full article

Android malware detection is increasingly important for mobile and edge security because malicious applications may compromise user privacy, device reliability, and sensitive service transactions. However, single-view static detection methods often provide limited semantic coverage and are sensitive to noisy or obfuscated code, while many deep learning models remain too heavy for resource-constrained deployment. To address these challenges, this paper proposes a lightweight Android malware detection framework based on SDAE-guided multi-view static feature fusion. The framework extracts three complementary static views, namely API calls, permission requests, and system components, from AndroidManifest.xml and classes.dex. These views are independently denoised and compressed by stacked denoising autoencoders, then aligned as the R, G, and B channels of a pseudo-RGB representation. A compact MicroNet-SE classifier with squeeze-and-excitation blocks is used to recalibrate the fused semantic channels and perform malware classification. Experiments on the CICMalDroid 2020 and CIC-AndMal2017 datasets show that the proposed framework achieves 99.01% accuracy, 99.15% precision, 98.99% recall, and 99.07% F1-score, with only 99.6 k parameters and a model size of 1.26 MB. After conversion to TensorFlow Lite, the MicroNet-SE classifier achieves average on-device inference latencies ranging from 1.85 ms to 2.16 ms on two real mobile devices. The model also maintains stable performance under synthetic feature perturbations and practical APK-level obfuscation settings. These findings suggest that combining multi-view static semantics with denoising-based representation learning can improve both detection robustness and deployment efficiency. Overall, the results indicate that the proposed framework provides an effective and lightweight static screening component for Android malware detection in resource-constrained mobile and edge environments. Full article

Real-time monitoring of endangered birds on edge AI hardware is constrained by a structural mismatch between modern lightweight detectors and mainstream NPU deployment toolchains. Recent attention-based detectors rely heavily on dynamic-shape operators that fall back to the host CPU on embedded NPUs, negating the advantages of lightweight architectures. To address this, we propose eBirdNet-Nano, a 1.05 M-parameter detector derived from YOLOv12n through a three-level NPU-friendly redesign: a static NPUConv block at the operator level, an NPU-C3k2 module together with an NPU-SE-Block at the module level, and a balanced 64-channel detection head at the head level. The resulting model achieves a 59% parameter reduction over YOLOv12n at only 5.8 GFLOPs while attaining an mAP@0.5 of 0.929 on a curated 24-species endangered-bird dataset collected in Macao. We further evaluate the model across four heterogeneous edge platforms—the Rockchip RK3588 (ARM + NPU), Kendryte K230 (RISC-V + KPU), Raspberry Pi 4B (pure ARM), and LicheePi 4A (pure RISC-V)—to characterize its behavior under distinct execution models. On the RK3588 NPU under INT8 quantization, eBirdNet-Nano delivers 13.83 ms inference latency and 26.76 ms end-to-end latency at 37.4 FPS, attaining the best parameter–latency balance and the highest parameter-normalized throughput (35.62 FPS/M) among six nano-scale YOLO variants, with an overall 3.53× end-to-end speedup over the YOLOv12n FP16 baseline that decomposes into a 2.97× architectural factor and a 1.19× quantization factor. Integrated into the EbirdEye field terminal, the same model sustains 23.5 ms thread-level end-to-end latency during live operation while supporting approximately 13.5 h of battery-powered runtime per charge. The proposed design offers a practical pathway toward deployable, low-power AI terminals for endangered-species conservation in resource-constrained field environments. Full article

Extreme and sudden temperature fluctuations observed as a result of global climate change increase the environmental pressure on energy transmission infrastructure. These meteorological changes significantly increase the risk of failure for porcelain insulators, which exhibit low thermal resistance and are susceptible to sudden arcing and surface deformations. In this study, a hybrid CNN–BiLSTM–SE architecture augmented with the Squeeze-and-Excitation attention mechanism is proposed using surface leakage current signals to diagnose healthy, cracked, and broken structural conditions in three-unit porcelain insulators. The SE block in the architecture dynamically rescales feature maps from CNN layers on a channel-by-channel basis. Thus, it highlights the signal characteristic that is dominant for fault diagnosis just before the BiLSTM units learn temporal dependencies. Leakage current data were obtained under an experimental setup at 60 kV for 15 different conditions covering all possible combinations of healthy, cracked, and broken insulator units. The raw signals were preprocessed with the Savitzky–Golay filter to suppress noise while preserving the diagnostic waveform morphology. 24 features covering time-domain statistics, frequency-domain spectral characteristics, and wavelet-domain energy components were extracted and used as model inputs. The CNN–BiLSTM–SE architecture achieved a classification accuracy of 93.83%, surpassing the standalone CNN (88.89%), BiLSTM (87.65%), and CNN–BiLSTM (91.36%) models, as well as classical machine-learning baselines (SVM: 87.65%, Random Forest: 90.12%, Boosted Trees: 87.65%). Full article

The Shiganghe tungsten–tin deposit and the Tiechang tin deposit are located in the northern part of the Baoshan Block, approximately 25 km apart. They were formed in the Late Cretaceous and Early Oligocene, respectively, and exhibit distinct deposit characteristics. This study presents a comparative analysis of in situ trace elements and S-Pb isotopes of pyrite from the two deposits to trace the sources of ore-forming fluids and materials and to further constrain the evolution of the metallogenic setting in the Baoshan Block. The trace element compositions of pyrite from the Shiganghe tungsten–tin deposit differ significantly from those of the Tiechang tin deposit. The former is relatively enriched in Se, Al, and W, while the latter contains relatively high concentrations of Cu, As, Sb, Pb, Bi, and Ag. The sulfur isotopic compositions of pyrite from the Shiganghe deposit are lower than those from the Tiechang deposit, with δ³⁴S_CDT values ranging from 2.11‰ to 6.86‰ and 6.74‰ to 8.72‰, respectively. The two deposits share consistent lead isotopic ratios, characterized by a mixture of upper crustal and orogenic belt sources. Therefore, the mineralization of the Shiganghe tungsten–tin deposit may be related to deep seated, concealed Late Cretaceous magmatism. Relatively oxidized and W-rich magmatic fluids contributed to tungsten–tin mineralization that characterized by a zoning pattern of “tungsten above and tin below”. In contrast, the Tiechang tin deposit is likely associated with coeval leucogranites present in the shear tectonic zone near Caojian. The ore-forming fluids underwent long distant migration with extensive crustal contamination, resulting in a relatively reduced hydrothermal system. Combined with the regional geological setting, the mineralization of the Shiganghe tungsten–tin deposit and the Tiechang tin deposit records the geological event from Neo-Tethys Ocean subduction to the collisional orogeny between the Indian and Eurasian blocks. The ore-forming materials originated from a mixture of upper crustal materials and orogenic belt materials. Full article

Intensive agricultural practices based on continuous monocropping and prolonged bare-soil fallows have contributed to soil degradation and loss of biological functioning. Replacing fallows with cover crops (CCs) is a promising strategy to restore soil quality, yet their legacy effects on rhizosphere fungal communities remain poorly understood. This study evaluated the legacy effects of Urochloa (syn. Brachiaria) brizantha cover cropping on rhizosphere fungal communities, as well as soil physicochemical and biological properties, in a degraded common bean system. A field experiment with a randomized complete block design included: bare fallow (BM), one (B1) or two (B2) CC cycles before bean, a perennial pasture (PB), and a pristine soil reference (PS). High-throughput sequencing showed that Urochloa-based treatments significantly shifted fungal community composition compared to BM, increasing saprotrophic and beneficial taxa (e.g., Mortierella, Penicillium, Coprinellus) and reducing potential pathogens such as Fusarium. These changes were associated with higher soil organic carbon, aggregate stability, microbial biomass, and enzyme activities, especially in B2 and PB. Indicator taxa identified by LEfSe were linked to organic matter decomposition and nutrient cycling. Multivariate analyses revealed strong associations between fungal community structure and soil properties. Overall, U. brizantha cover cropping induced measurable legacy effects, promoting soil biological recovery even after short-term implementation. Full article

Background: Electroencephalography (EEG) signals are widely used for monitoring anesthesia depth during surgery. Current commercial indicators are largely closed-source and may reflect dynamic changes with some delay. Methods: This study proposes a multitask deep learning model for continuous Bispectral Index (BIS) estimation, binary anesthesia-state classification, and prediction of transitions toward light anesthesia at different time intervals. Dual-channel EEG signals from 5471 surgical cases in the VitalDB dataset were divided into 60 s windows. Short-Time Fourier Transform (STFT) captured instantaneous frequency changes to transform the signal into a two-dimensional map. A ResNet-SE architecture incorporating Squeeze-and-Excitation blocks was used to identify EEG features associated with anesthesia depth. Results: A Mean Absolute Error of 3.27 and a Root Mean Square Error of 5.48 were obtained in anesthesia depth estimation. Light anesthesia classification achieved an AUC of 0.99 on the internal test set. Conclusions: The proposed multitask model enables the assessment of anesthesia depth and transitions toward light anesthesia using EEG signals. Full article

With the occurrence of global climate change and the depletion of agricultural water resources, there is a growing need to develop rapid, non-destructive, and autonomous plant health monitoring systems. As an economically valuable crop, Ocimum basilicum L. (basil) is sensitive to changes in water availability and may exhibit stress-related morphological variations under drought and over-irrigation conditions. However, due to the visual similarity of leaf symptoms under drought stress, waterlogging stress, and optimal irrigation conditions, accurately distinguishing these conditions remains challenging in practical applications. To address this challenge, this paper presents an attention-based dual-branch deep learning framework designed to extract both subtle leaf details and channel-related features from high-resolution plant images. By combining the Convolutional Block Attention Module (CBAM) and Squeeze-and-Excitation (SE) mechanism in a parallel structure, the proposed network improves the analysis of high-resolution images with an input size of 720 × 720 pixels. Under controlled environmental conditions, with ground-truth labels obtained using soil moisture sensor measurements, the proposed model was compared with eight deep learning architectures, including DenseNet121, InceptionV3, and VGG16. The proposed model achieved a hold-out evaluation accuracy of 99.54%, outperforming the second-best model, DenseNet121, which achieved 96.43%. In addition, the proposed model reached a class-specific precision value of 100% for the Drought Stress category and achieved an area under the receiver operating characteristic curve of 1.00 under the controlled experimental setting. Taylor Diagram analysis also indicated that the model closely preserved the variability pattern of the reference data. These results suggest that the proposed application-specific framework may support non-destructive basil water-stress detection under controlled conditions. After further validation with larger datasets, different cultivars, variable environmental conditions, and real-world agricultural scenarios, the proposed approach may contribute to precision irrigation management and sustainable agricultural production. The contribution of this study should be interpreted as an application-specific implementation and evaluation of complementary attention mechanisms for controlled-environment basil water-stress classification, rather than as the introduction of a fundamentally new deep learning methodology. Full article