Search Results (1,990)

Purpose: Gastrointestinal (GI) illness demands precise and efficient diagnostics, yet conventional approaches (e.g., endoscopy and histopathology) are time-consuming and prone to reader variability. This work presents GID-Xpert, a deep learning framework designed to improve feature learning, accuracy, and interpretability for GI disease classification. Methods: GID-Xpert integrates a hierarchical, multi-stage attention-driven mixture of experts with dynamic routing. The architecture couples spatial–channel attention mechanisms with specialized expert blocks; a routing module adaptively selects expert paths to enhance representation quality and reduce redundancy. The model is trained and evaluated on three benchmark datasets—WCEBleedGen, GastroEndoNet, and the King Abdulaziz University Hospital Capsule (KAUHC) dataset. Comparative experiments against state-of-the-art baselines and ablation studies (removing attention, expert blocks, and routing) are conducted to quantify the contribution of each component. Results: GID-Xpert achieves superior performance with 100% accuracy on WCEBleedGen, 99.98% on KAUHC, and 75.32% on GastroEndoNet. Comparative evaluations show consistent improvements over contemporary models, while ablations confirm the additive benefits of spatial–channel attention, expert specialization, and dynamic routing. The design also yields reduced computational cost and improved explanation quality via attention-driven reasoning. Conclusion: By unifying attention, expert specialization, and dynamic routing, GID-Xpert delivers accurate, computationally efficient, and more interpretable GI disease classification. These findings support GID-Xpert as a credible diagnostic aid and a strong foundation for future extensions toward broader GI pathologies and clinical integration. Full article

Sepsis-associated acute kidney injury (SA-AKI) is a prevalent and life-threatening complication in critically ill children, contributing to high mortality rates (up to 30%) and long-term renal dysfunction in pediatric intensive care units. This review synthesizes recent advances in the signalling pathways underlying SA-AKI, emphasizing pediatric-specific mechanisms, biomarkers, and therapeutic targets. This review covers inflammatory cascades via TLR/NF-κB leading to cytokine storms (IL-6, TNF-α); apoptosis and necrosis involving mitochondrial Bcl-2 dysregulation and OLFM4; and emerging processes like pyroptosis (NF-κB-mediated), metabolic reprogramming (choline deficiency and Nrf2-mitophagy), and novel routes such as cGAS-STING and TGF-β signalling. Biomarkers like urinary OLFM4, DKK3, NGAL, and serum suPAR, alanine, and Penkid enable early diagnosis and risk stratification, with models like PERSEVERE-II enhancing prognostic accuracy. Therapeutic strategies include fluid optimization, renal replacement therapies (CRRT, SLED-f), and pathway-targeted interventions such as choline supplementation, oXiris for cytokine removal, Humanin for immunomodulation, and investigational cGAS-STING inhibitors. Despite progress, challenges persist in translating animal models to pediatric trials and addressing heterogeneity. Integrating multi-omics and precision medicine holds promise for improving outcomes, underscoring the need for multicenter studies in children. Full article

During the Gas Metal Arc Welding (GMAW) process, intense arc light and dense fumes cause local overexposure in RGB images and data loss in point clouds, which severely compromises the extraction accuracy of circular closed-curve weld seams. To address this challenge, this paper proposes a multimodal fusion method for weld seam extraction under arc light and fume interference. The method begins by constructing a weld seam edge feature extraction (WSEF) module based on a synergistic fusion network, which achieves precise localization of the weld contour by coupling image arc light-removal and semantic segmentation tasks. Subsequently, an image-to-point cloud mapping-guided Local Point Cloud Feature extraction (LPCF) module was designed, incorporating the Shuffle Attention mechanism to enhance robustness against noise and occlusion. Building upon this, a cross-modal attention-driven multimodal feature fusion (MFF) module integrates 2D edge features with 3D structural information to generate a spatially consistent and detail-rich fused point cloud. Finally, a hierarchical trajectory reconstruction and smoothing method is employed to achieve high-precision reconstruction of the closed weld seam path. The experimental results demonstrate that under severe arc light and fume interference, the proposed method achieves a Root Mean Square Error below 0.6 mm, a maximum error not exceeding 1.2 mm, and a processing time under 5 s. Its performance significantly surpasses that of existing methods, showcasing excellent accuracy and robustness. Full article

To navigate dynamic environments safely, UAVs require accurate, real time onboard perception, which relies on ego motion compensation to separate self-induced motion from external dynamics and enable reliable obstacle detection. Traditional ego-motion compensation techniques are mainly based on optimization processes and may be computationally expensive for real-time applications or lack the precision needed to handle both rotational and translational movements, leading to issues such as misidentifying static elements as dynamic obstacles and generating false positives. In this paper, we propose a novel approach that integrates an event camera-based perception pipeline with an ego-motion compensation algorithm to accurately compensate for both rotational and translational UAV motion. An enhanced warping function, integrating IMU and depth data, is constructed to compensate camera motion based on real-time IMU data to remove ego motion from the asynchronous event stream, enhancing detection accuracy by reducing false positives and missed detections. On the compensated event stream, dynamic obstacles are detected by applying a motion aware adaptive threshold to the normalized mean timestamp image, with the threshold derived from the image’s spatial mean and standard deviation and adjusted by the UAV’s angular and linear velocities. Furthermore, in conjunction with a 3D Artificial Potential Field (APF) for obstacle avoidance, the proposed approach generates smooth, collision-free paths, addressing local minima issues through a rotational force component to ensure efficient UAV navigation in dynamic environments. The effectiveness of the proposed approach is validated through simulations, and its application for UAV navigation, safety, and efficiency in environments such as warehouses is demonstrated, where real-time response and precise obstacle avoidance are essential. Full article

Carbon-fiber-reinforced plastic (CFRP) machining by ultrashort-pulse lasers promises high precision but is limited due to the heterogeneous epoxy–carbon fiber structure, which creates heat-affected zones and variable kerf quality. This work investigates synthesized copper hydroxyphosphate as a laser-absorbing additive to improve femtosecond (1030 nm) laser ablation of CFRP. Copper hydroxyphosphate particles were synthesized hydrothermally and incorporated into an epoxy matrix to produce single-ply CFRP laminates. Square patterns (0.5 × 0.5 mm) were ablated with a pulse energy of 0.5–16 μJ. Then, ablated volumes were profiled and materials characterized by SEM and EDS. In neat epoxy the copper additive reduced optimum ablation efficiency and decreased penetration depth, while producing smoother, less porous surfaces. In contrast, CFRP with copper hydroxyphosphate showed increased efficiency and higher penetration depth. SEM and EDS analyses indicate more uniform matrix removal and retention of resin residues on fibers. These results suggest that copper hydroxyphosphate acts as a local energy absorber that trades volumetric removal for improved surface quality in epoxy and enhances uniformity and process stability in CFRP femtosecond laser machining. Full article

Water is an essential resource whose quality is threatened by emerging pollutants, including microplastics (MP), whose persistence, bioaccumulation capacity and ecotoxic potential pose a growing risk to ecosystems and human health. Wastewater treatment plants (WWTPs) have been identified as one of the main sources of these pollutants, as conventional treatments are insufficient to remove them completely. In response to this problem and with the aim of finding more efficient and sustainable solutions, a study has been carried out at WWTP with a pilot MP capture plant capable of detecting, quantifying and removing these particles from different wastewater sources with high precision and sustainability. This proposal represents a significant advance in the mitigation of invisible pollution, contributing to the protection of the environment and public health, achieving an efficiency of over 80% in the removal of plastic particles. This system not only addresses the challenge of environmental protection but also represents an unavoidable commitment to a healthier, more equitable, and sustainable development model for current and future generations, directly contributing to strategic action to advance the fulfillment of several Sustainable Development Goals (SDGs) promoted by the UN (SDG 3, SDG 6, SDG 12, SDG 14 and SDG 15). Full article

Introduction Artificial intelligence (AI) has been influencing the way oncology has been practiced. Major issues constituting a bottleneck are the lack of data for training purposes, confidentiality preventing development, or the absence of transparency in clarifying how models operate to generate decisions. Novel Models With explainable AI, trust and utilization barriers among clinicians, researchers, and patients can be removed. With the implementation of federated learning, multiple institutions could contribute to crucial dataset’s learning information. Precise diagnosis and prescription of the right drug are essential in preventing unnecessary life losses, and economic burden to the underling system. Focus This review focuses on new AI models that could make medical diagnosis more precise, quicker and convenient, as well as help with the discovery of new drugs and better selection of certain cancer therapies, in particular for those drugs whose results have been inconsistent across different groups of patients. We then speculate on the transformative role AI could play in predicting ADC therapy efficacy. This would ultimately bestow the medical field of an impeccable selection tool. Such colossal methodology, coupled with seeming monitoring of treatment and recovery, may be granting remedies that have been so longed, and deemed necessary. Full article

Accurate segmentation of abdominal aortic aneurysm (AAA) from computed tomography (CT) images is critical for early diagnosis and treatment planning of vascular diseases. However, noise in CT images obscures vessel boundaries, reducing segmentation accuracy. U-Net is widely used for medical image segmentation, where noise removal is critical. This study applied various denoising filters for U-Net segmentation and classified the severity of segmented AAA images to evaluate accuracy. Poisson–Gaussian noise was added to AAA CT images, and then average, median, Wiener, and median-modified Wiener filters (MMWF) were applied. U-Net-based segmentation was performed, and the segmentation accuracy of the output images obtained per filter was quantitatively assessed. Furthermore, the Hough circle algorithm was applied to the segmented images for diameter measurement, enabling severity classification and evaluation of classification accuracy. MMWF application improved the Matthews correlation coefficient, Dice score, Jaccard coefficient, and mean surface distance by 31.09%, 34.25%, 53.99%, and 3.70%, respectively, compared with images with added noise. Moreover, classification based on the output images obtained after MMWF application demonstrated the highest accuracy, with sensitivity, precision, and accuracy reaching 100%. Thus, U-Net-based segmentation yields more accurate results when images are processed with the MMWF and analyzed using the Hough circle algorithm. Full article

Broadcast-style downlinks (e.g., PONs and satellites) expose physical waveforms despite transport-layer cryptography, motivating physical-layer encryption (PLE). Digital chaotic encryption is appealing for its noise-like spectra, sensitivity, and DSP-friendly implementation, but in low-CSPR KK-SSB systems, common embeddings disrupt minimum-phase requirements and raise PAPR/SSBI near 1 dB CSPR, while finite-precision effects can leak correlation after KK reconstruction. We bridge this gap by integrating 3D Lorenz-based PLE into our low-CSPR KK-SSB receiver. A KK-compatible embedding applies a Lorenz-driven XOR mapping to I/Q bitstreams before PAM4-to-16QAM modulation, preserving the minimum phase and avoiding spectral zeros. Co-design of chaotic strength and subband usage with the KK SSBI-suppression method maintains SSBI mitigation with negligible PAPR growth. We further adopt digitization settings and fractional-digit-parity-based key derivation to suppress short periods and remove key-revealing synchronization cues. Experiments demonstrate a 10⁹¹ key space without degrading transmission quality, enabling secure, key-concealed operation on shared downlinks and offering a practical path for chaotic PLE in near-minimum-CSPR SSB systems. Full article

Precision dicing with diamond wheels is a key technology in semiconductor dicing, integrated circuit manufacturing, aerospace, and other fields, owing to its high precision, high efficiency, and broad material applicability. As a critical processing stage, a comprehensive analysis of dicing technologies is essential for improving the machining quality of hard-and-brittle optoelectronic materials. This paper reviews the core principles of precision diamond wheel dicing, including dicing processes and blade preparation methods. Specifically, it examines the dicing mechanisms of composite and multi-mode dicing processes, demonstrating their efficacy in reducing defects inherent to single-mode approaches. The review also examines diverse preparation methods for dicing blades, such as metal binder sintering and roll forming. Furthermore, the roles of machine vision and servo control systems are detailed, illustrating how advanced algorithms facilitate precise feature recognition and scribe line control. A systematic analysis of key components in grinding wheel dicer is also conducted to reduce dicing deviation. Additionally, the review introduces models for tool wear detection and discusses material removal mechanisms. The influence of critical process parameters—such as spindle speed, feed rate, and dicing depth—on dicing quality and kerf width is also analyzed. Finally, the paper outlines future prospects and provides recommendations for advancing key technologies in precision dicing, offering a valuable reference for subsequent research. Full article

Accurate prediction of concrete compressive strength is essential for ensuring structural safety in civil engineering, particularly in road and bridge construction, where inadequate strength can lead to deformation, cracking, or collapse. Traditional non-destructive testing (NDT) methods, such as the Rebound Hammer Test, estimate strength using regression-based formulas fitted with measurement data; however, these formulas, typically optimized via the least squares method, are highly sensitive to initial parameter settings and exhibit low robustness, especially for nonlinear relationships. Meanwhile, AI-based models, such as neural networks, require extensive datasets for training, which poses a significant challenge in real-world engineering scenarios with limited or unevenly distributed data. To address these issues, this study proposes a gradient-boosting artificial bee colony (GB-ABC) algorithm for robust regression curve fitting. The method integrates two novel mechanisms: gradient descent to accelerate convergence and prevent entrapment in local optima, and a point density-weighted strategy using Gaussian Kernel Density Estimation (GKDE) to assign higher weights to sparse data regions, enhancing adaptability to field data irregularities without necessitating large datasets. Following data preprocessing with Local Outlier Factor (LOF) to remove outliers, validation on 600 real-world samples demonstrates that GB-ABC outperforms conventional methods by minimizing mean relative error rate (RER) and achieving precise rebound-strength correlations. These advancements establish GB-ABC as a practical, data-efficient solution for on-site concrete strength estimation. Full article

Multi Jet Fusion (MJF) is a leading technology for producing functional polymer parts. However, it still faces challenges with dimensional accuracy and removing unfused powder from complex internal geometries. First, dimensional accuracy was mapped by producing 45 identical PA12 specimens on an HP MJF 4200 printer in a 5 × 9 layout across five vertical layers. The analysis revealed a consistent pattern: parts located in the central positions of the build volume exhibited the poorest accuracy, while those near the perimeter were the most precise, regardless of their vertical height. This spatial variation is attributed to non-uniform thermal control from the printer’s adaptive lamp–thermal camera system. Second, the limits of powder removal from closed body-centered cubic (BCC) lattice structures were quantified. Using sandblasting and X-ray inspection, a strong inverse relationship was found between a lattice’s relative density and the maximum thickness that could be thoroughly cleaned of powder. For example, low-density structures (ρ = 0.07) could be cleaned up to five layers deep, whereas high-density structures (ρ = 0.39–0.47) were limited to only 1.5–1.7 layers. These findings offer actionable guidelines for optimizing part placement and designing internal lattice structures for MJF technology. The key findings are the spatial variation in dimensional accuracy in MJF printing, where the central parts are the least accurate and perimeter parts are the most precise, and the inverse relationship between a lattice’s relative density (ρ) and cleanable thickness. Specifically, low-density structures (ρ = 0.07) could be thoroughly cleaned up to five layers, while high-density ones (ρ = 0.39–0.47) were limited to approximately 1.5–1.7 layers. The layer thickness was a pre-designed parameter (2, 3, 4, and 5 layers), and powder removal was supported by using automated sandblasting followed by verification via industrial X-ray imaging. Full article

Autophagy (cellular self-eating) is a tightly regulated catabolic process of eukaryotic cells during which parts of the cytoplasm are sequestered and subsequently delivered into lysosomes for degradation by acidic hydrolases. This process is central to maintaining cellular homeostasis, the removal of aged or damaged organelles, and the elimination of intracellular pathogens. The nematode Caenorhabditis elegans has proven to be a powerful genetic model for investigating the regulation and mechanism of autophagy. To date, the fluorescent autophagy reporters developed in this organism have predominantly relied on multi-copy, randomly integrated transgenes. As a result, the interpretation of autophagy dynamics in these models has required considerable caution due to possible overexpression artifacts and positional effects. In addition, starvation-induced autophagy has not been characterized in detail using these reporters. Here, we describe the development of two endogenous autophagy reporters, gfp::mCherry::lgg-1/atg-8 and gfp::atg-5, both inserted precisely into their endogenous genomic loci. We demonstrate that these single-copy reporters reliably track distinct stages of the autophagic process. Using these tools, we reveal that (i) the transition from the earliest phagophore to the mature autolysosome is an exceptionally rapid event because the vast majority of the detected fluorescent signals are autolysosome-specific, (ii) starvation triggers autophagy only after a measurable lag phase rather than immediately, and (iii) the regulation of starvation-induced autophagy depends on the actual life stage, and prevents excessive flux that could otherwise compromise cellular survival. We anticipate that these newly developed reporter strains will provide refined opportunities to further dissect the physiological and pathological roles of autophagy in vivo. Full article

Malignant gliomas remain among the most formidable challenges in neuro-oncology, given their high morbidity and rising incidence worldwide. Surgical resection represents the cornerstone of treatment, typically followed by adjuvant radiotherapy and chemotherapy. Achieving maximal safe resection, however, requires advanced intraoperative guidance. A range of adjuncts are currently employed, including 5-aminolevulinic acid (5-ALA), intraoperative ultrasound, computed tomography (iCT), and intraoperative magnetic resonance imaging (iMRI). More recently, an emerging technique—virtual MRI (vMRI)—has been developed, fusing intraoperative CT with preoperative high-resolution MRI to provide real-time, MRI-like updates of brain anatomy. Beyond imaging, tumour removal itself induces reorganization of eloquent brain networks, underscoring the critical need for precision tools that balance oncological control with preservation of neurological function. In this narrative review, we highlight and synthesize the evolving armamentarium of intraoperative technologies shaping the future of precision neuro-oncology. Full article

The development of high-performance MEMS-based eye trackers, crucial for next-generation medical diagnostics and human–computer interfaces, is often hampered by the mechanical instability and time-consuming recalibration of physical prototypes. To address this bottleneck, we present the development and rigorous validation of a high-fidelity digital twin (DT) designed to accelerate the design–test–refine cycle. We conducted a comparative study of a physical MEMS scanning system and its corresponding digital twin using a USAF 1951 test target under both static and dynamic conditions. Our analysis reveals that the DT accurately replicates the physical system’s behavior, showing a geometric discrepancy of <30 µm and a matching feature shift (1 µm error) caused by tracking dynamics. Crucially, the DT effectively removes mechanical vibration artifacts, enabling the precise analysis of system parameters in a controlled virtual environment. The validated model was then used to develop a pupil detection algorithm that achieved an accuracy of 1.80 arc minutes, a result that surpasses the performance of a widely used commercial system in our comparative tests. This work establishes a validated methodology for using digital twins in the rapid prototyping and optimization of complex optical systems, paving the way for faster development of critical healthcare technologies. Full article