Search Results (74)

Small-object detection in UAV remote sensing faces common challenges such as tiny target size, blurred features, and severe background interference. Furthermore, single imaging modalities exhibit limited representation capability in complex environments. To address these issues, this paper proposes CTU-YOLO, a UAV-based small-object detection algorithm built upon cross-modal feature fusion and lightweight upsampling. The algorithm incorporates a dynamic and adaptive cross-modal feature fusion (DCFF) module, which achieves efficient feature alignment and fusion by combining frequency-domain analysis with convolutional operations. Additionally, a lightweight upsampling module (LUS) is introduced, integrating dynamic sampling and depthwise separable convolution to enhance the recovery of fine details for small objects. Experiments on the DroneVehicle and LLVIP datasets demonstrate that CTU-YOLO achieves 73.9% mAP on DroneVehicle and 96.9% AP on LLVIP, outperforming existing mainstream methods. Meanwhile, the model possesses only 4.2 MB parameters and 13.8 GFLOPs computational cost, with inference speeds reaching 129.9 FPS on DroneVehicle and 135.1 FPS on LLVIP. This exhibits an excellent lightweight design and real-time performance while maintaining high accuracy. Ablation studies confirm that both the DCFF and LUS modules contribute significantly to performance gains. Visualization analysis further indicates that the proposed method can accurately preserve the structure of small objects even under nighttime, low-light, and multi-scale background conditions, demonstrating strong robustness. Full article

The deep learning paradigm is progressively shifting from non-smooth activation functions, exemplified by ReLU, to smoother alternatives such as GELU and SiLU. This transition is motivated by the fact that non-differentiability introduces challenges for gradient-based optimization, while an expanding body of research demonstrates that smooth activations yield superior convergence, improved generalization, and enhanced training stability. A central challenge, however, is how to systematically transform widely used non-smooth functions into smooth counterparts that preserve their proven representational strengths while improving differentiability and computational efficiency. To address this, we propose a general activation smoothing framework grounded in mollification theory. Leveraging the Epanechnikov kernel, the framework achieves statistical optimality and computational tractability, thereby combining theoretical rigor with practical utility. Within this framework, we introduce Smoothed ReLU (S-ReLU), a novel second-order continuously differentiable (C²) activation derived from ReLU that inherits its favorable properties while mitigating inherent drawbacks. Extensive experiments on CIFAR-10, CIFAR-100, and ImageNet-1K with Vision Transformers and ConvNeXt consistently demonstrate the superior performance of S-ReLU over existing ReLU variants. Beyond computer vision, large-scale fine-tuning experiments on language models further show that S-ReLU surpasses GELU, underscoring its broad applicability across both vision and language domains and its potential to enhance stability and scalability. Full article

To address the challenges of poor adaptability to spatial heterogeneity, easy breakage of amplitude–phase coupling relationships, and insufficient physical consistency in complex optical wavefield reconstruction, this paper proposes the DdONN-PINNs hybrid framework. Focused on preserving the intrinsic symmetries of wave physics, the framework achieves deep integration of optical neural networks and physics-informed information. Centered on an architecture of “SIREN shared encoding–domain-specific output”, it utilizes the periodic activation property of SIREN encoders to maintain the spatial symmetry of wavefield distribution, incorporates learnable Fourier diffraction layers to model physical propagation processes, and adopts native complex-domain modeling to avoid splitting the real and imaginary parts of complex amplitudes—effectively adapting to spatial heterogeneity while fully preserving amplitude-phase coupling in wavefields. Validated on rogue wavefields governed by the Nonlinear Schrödinger Equation (NLSE), experimental results demonstrate that DdONN-PINNs achieve an amplitude Mean Squared Error (MSE) of $2.94\times {10}^{-3}$ and a phase MSE of $5.86\times {10}^{-4}$, outperforming non-domain-decomposed models and ReLU-activated variants significantly. Robustness analysis shows stable reconstruction performance even at a noise level of $\sigma =0.1$. This framework provides a balanced solution for wavefield reconstruction that integrates precision, physical interpretability, and robustness, with potential applications in fiber-optic communication and ocean optics. Full article

Stomach ulcers, a common type of gastrointestinal (GI) disease, pose serious health risks if not diagnosed and treated at an early stage. Therefore, in this research, a method is proposed based on two deep learning models for classification and segmentation. The classification model is based on Convolutional Neural Networks (CNN) and incorporates Q-learning to achieve learning stability and decision accuracy through reinforcement-based feedback. In this model, input images are passed through a custom CNN model comprising seven layers, including convolutional, ReLU, max pooling, flattening, and fully connected layers, for feature extraction. Furthermore, the agent selects an action (class) for each input and receives a +1 reward for a correct prediction and −1 for an incorrect one. The Q-table stores a mapping between image features (states) and class predictions (actions), and is updated at each step based on the reward using the Q-learning update rule. This process runs over 1000 episodes and utilizes Q-learning parameters (α = 0.1, γ = 0.6, ϵ = 0.1) to help the agent learn an optimal classification strategy. After training, the agent is evaluated on the test data using only its learned policy. The classified ulcer images are passed to the proposed attention-based U-Net model to segment the lesion regions. The model contains an encoder, a decoder, and attention layers. The encoder block extracts features through pooling and convolution layers, while the decoder block up-samples the features and reconstructs the segmentation map. Similarly, the attention block is used to highlight the important features obtained from the encoder block before passing them to the decoder block, helping the model focus on relevant spatial information. The model is trained using the selected hyperparameters, including an 8-batch size, the Adam optimizer, and 50 epochs. The performance of the models is evaluated on Kvasir, Nerthus, CVC-ClinicDB, and a private POF dataset. The classification framework provides 99.08% accuracy on Kvasir and 100% accuracy on Nerthus. In contrast, the segmentation framework yields 98.09% accuracy on Kvasir, 99.77% accuracy on Nerthus, 98.49% accuracy on CVC-ClinicDB, and 99.13% accuracy on the private dataset. The achieved results are superior to those of previous methods published in this domain. Full article

In the domain of automatic parking systems, parking space detection and localization represent fundamental challenges that must be addressed. As a core research focus within the field of intelligent automatic parking, they constitute the essential prerequisite for the realization of fully autonomous parking. Accurate and effective detection of parking spaces is still the core problem that needs to be solved in automatic parking systems. In this study, building upon existing public parking space datasets, a comprehensive panoramic parking space dataset named PSEX (Parking Slot Extended) with complex environmental diversity was constructed by integrating the concept of GAN (Generative Adversarial Network)-based image style transfer. Meanwhile, an improved algorithm based on PP-Yoloe (Paddle-Paddle Yoloe) is used to detect the state (free or occupied) and angle (T-shaped or L-shaped) of the parking space in real-time. For the many and small labels of the parking space, the ResSpp in it is replaced by the ResSimSppf module, the SimSppf structure is introduced at the neck end, and Silu is replaced by Relu in the basic structure of the CBS (Conv-BN-SiLU), and finally an auxiliary detector head is added at the prediction head. Experimental results show that the proposed SimSppf_mepre-Yoloe model achieves an average improvement of 4.5% in mAP50 and 2.95% in mAP50:95 over the baseline PP-Yoloe across various parking space detection tasks. In terms of efficiency, the model maintains comparable inference latency with the baseline, reaching up to 33.7 FPS on the Jetson AGX Xavier platform under TensorRT optimization. And the improved enhancement algorithm can greatly enrich the diversity of parking space data. These results demonstrate that the proposed model achieves a better balance between detection accuracy and real-time performance, making it suitable for deployment in intelligent vehicle and robotic perception systems. Full article

Flaxseed polysaccharides (FLP) are bioactive macromolecules with valuable functional properties and applications in the food, pharmaceutical, and packaging industries. This study focused on obtaining high-purity pectin from flaxseed cake using sustainable extraction with natural deep eutectic solvents (NADES) based on choline chloride (ChCl) and citric acid (CA) The ChCl/CA system (1:1) resulted in the LU3 extract, which provided the best outcome, yielding the highest pectin recovery (36.88 mg/g), elevated uronic acid content (30.33% of sample; 68.15% of saccharides), and the lowest protein contamination (11.46%), confirming superior pectin purity. Structural (UV-Vis, FT-IR, GC-MS, GPC, LH-20) identified homogalacturonan with xylogalacturonan domains (53% DM) and a molecular weight range of 14–500 × 10³ g/mol. Morphological and physicochemical characterization, including SEM/EDS imaging, zeta potential analysis, and rheological measurements, revealed that LU3 is an anionic, heterogeneous biopolymer exhibiting pH-dependent charge behavior. These properties underscore its potential as a safe and effective material for bio-industrial applications. Overall, the study demonstrates that NADES provide an eco-friendly and efficient medium for extracting high-quality pectin from flaxseed cake, offering a sustainable strategy for the valorization of flaxseed polysaccharides in bio-based products. Full article

To improve the performance of image super-resolution reconstruction, this paper optimizes the classical SRGAN model architecture. The original SRResNet is replaced with the EDSR network as the generator, which effectively enhances the ability to restore image details. To address the issue of insufficient multi-scale feature extraction in SRGAN during image reconstruction, an LSK attention mechanism is introduced into the generator. By fusing features from different receptive fields through parallel multi-scale convolution kernels, the model improves its ability to capture key details. To mitigate the instability and overfitting problems in the discriminator training, the Mish activation function is used instead of LeakyReLU to improve gradient flow, and a Dropout layer is introduced to enhance the discriminator’s generalization ability, preventing overfitting to the generator. Additionally, a staged training strategy is employed during adversarial training. Experimental results show that the improved model effectively enhances image reconstruction quality while maintaining low complexity. The generated results exhibit clearer details and more natural visual effects. On the public datasets Set5, Set14, and BSD100, compared to the original SRGAN, the PSNR and SSIM metrics improved by 13.4% and 5.9%, 9.9% and 6.0%, and 6.8% and 5.8%, respectively, significantly enhancing the reconstruction of super-resolution images, achieving more refined and realistic image quality improvement. The model also demonstrates stronger generalization ability on complex cross-domain data, such as remote sensing images and medical images. The improved model achieves higher-quality image reconstruction and more natural visual effects while maintaining moderate computational overhead, validating the effectiveness of the proposed improvements. Full article

The southern Great Xing’an Range is located in the overlap zone of the Paleo-Asian Ocean metallogenic domain and the Circum-Pacific metallogenic domain. It hosts numerous Sn-polymetallic deposits, such as Weilasituo, Bianjiadayuan, Huanggang, and Dajing, and witnessed multiple episodes of magmatism during the Late Mesozoic. The study area is situated within the Huanggangliang-Ganzhuermiao metallogenic belt in the southern Great Xing’an Range. The region has witnessed extensive magmatism, with Mesozoic magmatic activities being particularly closely linked to regional mineralization. We present petrographic, zircon U-Pb chronological, lithogeochemical, and Lu-Hf isotopic analyses of the Yexilinhundi granites. The results indicate that the granite porphyry and granodiorite were emplaced during the Late Jurassic. Both rocks exhibit high SiO₂, K₂O + Na₂O, differentiation index (DI), and 10,000 Ga/Al ratios, coupled with low MgO contents. They show distinct fractionation between light and heavy rare earth elements (LREEs and HREEs), exhibit Eu anomalies, and have low whole-rock zircon saturation temperatures (T_zr), collectively demonstrating characteristics of highly fractionated I-type granites. The εHf(t) values of the granites range from 0.600 to 9.14, with young two-stage model ages (T_DM2 = 616.0~1158 Ma), indicating that the magmatic source originated from partial melting of Mesoproterozoic-Neoproterozoic juvenile crust. This study proposes that the granites formed in a post-collisional/post-orogenic extensional setting associated with the subduction of the Mongol-Okhotsk Ocean, providing a scientific basis for understanding the relationship between the formation of Sn-polymetallic deposits and granitic magmatic evolution in the study area. Full article

Plant triterpenoids are structurally diverse specialized metabolites with significant ecological, medicinal, and agricultural importance. Oxidosqualene cyclases (OSCs) catalyze the crucial cyclization step in triterpenoid biosynthesis, generating the fundamental carbon skeletons that determine their structural diversity and biological functions. Genome-wide identification of OSC genes was performed using bioinformatics tools, including HMMER and BLASTP, followed by phylogenetic analysis, gene structure analysis, conserved domain and motifs identification, cis-regulatory element prediction, protein–protein interaction analysis, and expression profiling using publicly available transcriptome data from UV-B treated A. annua six-week-old seedlings. We identified 24 AaOSC genes, classified into CAS, LAS, LUS, and unknown subfamilies. Phylogenetic analysis revealed evolutionary relationships with OSCs from other plant species. Gene structure analysis showed variations in exon–intron organization. Promoter analysis identified cis-regulatory elements related to light responsiveness, plant growth and development, hormone signaling, and stress response. Expression profiling revealed differential expression patterns of AaOSC genes under UV-B irradiation. This genome-wide characterization provides insights into the evolution and functional diversification of the OSC gene family in A. annua. The identified AaOSC genes and their regulatory elements lay the foundation for future studies aimed at manipulating triterpenoid biosynthesis for medicinal and biotechnological applications, particularly focusing on enhancing stress tolerance and artemisinin production. Full article

Physical-layer (PHY) security is widely used as an effective method for ensuring secure wireless communications. However, when the legitimate user (LU) and the eavesdropper (Eve) are in close proximity, the channel coupling can significantly degrade the secure performance of PHY. Frequency diverse array (FDA) technique addresses channel coupling issues by introducing frequency offsets among array elements. However, FDA’s ability to secure communication relies mainly on frequency domain characteristics, lacking the spatial degrees of freedom. The recently proposed movable antenna (MA) technology serves as an effective approach to overcome this limitation. It offers the flexibility to adjust antenna positions dynamically, thereby further decoupling the channels between LU and Eve. In this paper, we propose a novel MA-FDA approach, which offers a comprehensive solution for enhancing PHY security. We aim to maximize the achievable secrecy rate through the joint optimization of all antenna positions at the base station (BS), FDA frequency offsets, and beamformer, subject to the predefined regions for antenna positions, frequency offsets range, and energy constraints. To solve this non-convex optimization problem, which involves highly coupled variables, the alternating optimization (AO) method is employed to cyclically update the parameters, with the projected gradient ascent (PGA) method and block successive upper-bound minimization (BSUM) method being employed to tackle the challenging subproblems. Simulation results demonstrate that the MA-FDA approach can achieve a higher secrecy rate compared to the conventional phased array (PA) or fixed-position antenna (FPA) schemes. Full article

U-Pb and Lu-Hf isotopes, by inductively coupled plasma mass spectrometry and laser ablation (ICP-MS-LA), are reported in zircon grains from the Peixe Alkaline Suite. This unit comprises alkaline rocks such as syenites with nepheline, albite-oligoclase-biotite, and pegmatitic bodies. The zircon grain was imaged by cathodoluminescence (CL), which allowed the characterization of features within the crystal. These features comprise complex zone crosscuts, showing the existence of pulses that caused the intrusion of isotopically younger phases into the interior of the grain on a millimetric scale. The U-Pb results suggest a metamorphic event with Pb loss at 579 ± 3 Ma. They can be interpreted because of the collisional regional event of the Brasilia Orogen (Mara Rosa Orogeny). A second age grouping at 548 ± 2.5 Ma (MSWD = 8), obtained in areas with high luminescence fading laterally to oscillatory zoned domains with variations in the abundance of isotopes, is 33 Ma younger, demonstrating a rejuvenation of these areas through Pb loss. It is interpreted here as a second metamorphic event related to a collisional event (Santa Terezinha de Goiás arc). The Lu-Hf results for these areas indicate ƐHf values between −10 and −17, suggesting the existence of magmatic isotopic rework in a crustal environment. Full article

This work explores the deposition of hexagonal (h-) LuMnO₃ thin films in the P6₃cm phase and investigates the conditions under which the synergy of ferroelectric and photoactive properties, can be achieved to confirm the potential of this material for applications in the development of next-generation photovoltaic devices. Single-phase h-LuMnO₃ was successfully deposited on different substrates, and the thermal stability of the material was confirmed by Micro-Raman spectroscopy analysis from 77 to 850 K, revealing the suitable ferro- to para-electric transition near 760 K. Optical measurements confirm the relatively narrow band gap at 1.5 eV, which corresponds to the h-LuMnO₃ system. The presence of domain structures and the signature of hysteresis loops consistent with ferroelectric behaviour were confirmed by piezoresponse force microscopy. In addition, light-dependent photocurrent measurements revealed the photoactive sensitivity of the material. Full article

Improving user-level energy efficiency is critical for reducing the load on the power grid and addressing the challenges created by tight power balance when operating domestic air conditioning equipment under time-of-use (ToU) pricing. This paper presents a data-driven control method for HVAC (heating, ventilation, and air conditioning) systems that is based on model predictive control (MPC) and takes ToU electricity pricing into account. To describe building thermal dynamics, a multi-layer neural network is constructed using time-delayed embedding, with the rectified linear unit (ReLU) serving as the activation function for hidden layers. Using this piecewise affine approximation, an optimization model is developed within a receding horizon control framework, integrating the data-driven model and transforming it into a mixed-integer linear programming issue for efficient problem solving. Furthermore, this research suggests a hybrid optimization model for integrating air conditioning systems and battery energy storage systems. By employing a rolling time-domain control method, the proposed model minimizes the frequency of switching between charging and discharging states of the battery energy storage system, improving system reliability and efficiency. An Internet of Things (IoT)-based home energy management system is developed and validated in a real laboratory environment, complemented by a distributed integration solution for the energy management monitoring platform and other essential components. The simulation results and field measurements demonstrate the system’s effectiveness, revealing discernible pre-cooling and pre-charging behaviors prior to peak electricity pricing periods. This cooperative economic operation reduces electricity expenses by 13% compared to standalone operation. Full article

Sickle cell disease (SCD) presents complex clinical manifestations influenced by genetic, social, environmental, and healthcare access factors as well as socioeconomic status. In this context, sickle cell leg ulcers (SLUs) are a debilitating complication of SCD. We aimed to describe sociodemographic data and evaluate the quality of life (QoL) of SCD patients with and without SLUs. We conducted a cross-sectional study including 13 SCD patients with SLUs and 42 without LUs. Clinical data were obtained by reviewing the medical records, and QoL was assessed with the WHOQOL-BREF questionnaire. Our cohort of patients had a mean age of 34.9 years, with 52.8% male, 52.8% identifying as black, and 41.7% identifying as brown. Most had low income, incomplete education, and high unemployment rates. The social habits and relationships of SCD patients showed varying levels of friendship and family closeness, and the majority of SLU+ patients did not practice sports. We failed to find statistical differences in the WHOQOL-BREF domains between SLU+ and SLU− patients. However, higher income and employment status were associated with improved WHOQOL-BREF domain scores in SCD patients, while vaso-occlusive episodes and female gender were linked to lower scores. Our data reinforce the sociodemographic characteristics of SCD. The physical domain was associated with income, occupation, and vaso-occlusion. The psychological domain was associated with income and occupation. The social relationship domain was associated with occupation and female gender. The environmental domain was associated with vaso-occlusion. The WHOQOL-BREF is a reliable tool to measure QoL in SCD. Full article

Electroencephalography (EEG) is a non-invasive technique widely used in neuroscience to diagnose neural disorders and analyse brain activity. However, ocular and myogenic artifacts from eye movements and facial muscle activity often contaminate EEG signals, compromising signal analysis accuracy. While deep learning models are a popular choice for denoising EEG signals, most focus on removing either ocular or myogenic artifacts independently. This paper introduces a novel EEG denoising model capable of handling the simultaneous occurrence of both artifacts. The model uses convolutional layers to extract spatial features and a fully connected layer to reconstruct clean signals from learned features. The model integrates the Adam optimiser, average pooling, and ReLU activation to effectively capture and restore clean EEG signals. It demonstrates superior performance, achieving low training and validation losses with a significantly reduced $RRMSE$ value of 0.35 in both the temporal and spectral domains. A high cross-correlation coefficient of 0.94 with ground-truth EEG signals confirms the model’s fidelity. Compared to the existing architectures and models (FPN, UNet, MCGUNet, LinkNet, MultiResUNet3+, Simple CNN, Complex CNN) across a range of signal-to-noise ratio values, the model shows superior performance for artifact removal. It also mitigates overfitting, underscoring its robustness in artifact suppression. Full article