Search Results (315)

Segmentation of white blood cells is critical for a wide range of applications. It aims to identify and isolate individual white blood cells from medical images, enabling accurate diagnosis and monitoring of diseases. In the last decade, many researchers have focused on this task using U-Net, one of the most used deep learning architectures. To further enhance segmentation accuracy and robustness, recent advances have explored the combination of U-Net with other techniques, such as attention mechanisms and aggregation techniques. However, a common challenge in white blood cell image segmentation is the similarity between the cells’ cytoplasm and other surrounding blood components, which often leads to inaccurate or incomplete segmentation due to difficulties in distinguishing low-contrast or subtle boundaries, leaving a significant gap for improvement. In this paper, we propose GAAD-U-Net, a novel architecture that integrates attention-augmented convolutions to better capture ambiguous boundaries and complex structures such as overlapping cells and low-contrast regions, followed by a gating mechanism to further suppress irrelevant feature information. These two key components are integrated in the Double U-Net base architecture. Our model achieves state-of-the-art performance on white blood cell benchmark datasets, with a 3.4% Dice score coefficient (DSC) improvement specifically on the SegPC-2021 dataset. The proposed model achieves superior performance as measured by mean the intersection over union (IoU) and DSC, with notably strong segmentation performance even for difficult images. Full article

Wide-Bandgap (WBG) semiconductors—silicon carbide (SiC) and gallium nitride (GaN)— enable high-power-density conversion, but performance is limited by where heat is generated and how it is removed. This review links device-level loss mechanisms (conduction and switching, including output-capacitance hysteresis and dynamic on-resistance) to structure-driven hot spots within the ultra-thin (tens of nanometers) two-dimensional electron gas (2DEG) channel of GaN HEMTs and to thermal boundary resistance at layer interfaces. We compare wire-bondless package concepts—double-sided cooling, embedded packaging, and interleaved planar layouts—and survey system-level cooling that shortens the conduction path and raises heat-transfer coefficients. The impact on reliability is discussed using temperature-sensitive electrical parameters (e.g., on-state VDS, threshold voltage, drain leakage, di/dt, and gate current) for real-time junction-temperature estimation and compact electro-thermal RC models for remaining-useful-life prediction. Evidence from recent literature points to interface resistance in GaN-on-SiC as a primary bottleneck, while near-junction cooling and advanced packages are effective mitigations. We argue for integrated co-design—devices, packaging, electromagnetic interference (EMI)-aware layout, and cooling—together with interface engineering and health monitoring to deliver reliable, high-density WBG systems. Full article

Amid global efforts to reach carbon neutrality, quantifying the cradle-to-gate carbon footprint of panel kitchen cabinets is vital for the transformation of China’s furniture industry to low carbon emissions. This study aims to quantify and compare the cradle-to-gate carbon footprints of three L-shaped panel cabinets made of different materials and to identify the most effective carbon reduction strategies for the Chinese furniture industry. The emission factor method proposed by the Intergovernmental Panel on Climate Change (IPCC) was utilized. The results revealed significant differences in the carbon footprints among the three cabinet products. Specifically, Product A, featuring a DuPont stone countertop from the United States and domestically produced double-sided decorative door panels, exhibited the highest carbon footprint which was 998.5 kgCO₂eq. Product B, with an Italian natural marble countertop and single-sided acrylic door panels, had the lowest carbon footprint which was 610.7 kgCO₂eq. The carbon footprints indicated that key stages such as cabinet bodies, countertops, hardware, and cabinet doors were substantial contributors. Raw material production and processing emerged as the primary sources of carbon emissions, with countertop transportation also contributing significantly. Based on the results, this paper proposed several carbon reduction suggestions. These include optimizing material selection, enhancing energy efficiency in raw material production and processing, optimizing transportation methods, emphasizing the carbon reduction potential of hardware components, and strengthening carbon footprint monitoring and management. Full article

Gamma-aminobutyric acid type A receptors (GABA_ARs) are pentameric ligand-gated ion channels essential for inhibitory neurotransmission in the central nervous system. Subtype-specific expression patterns of GABA_AR subunits underlie their diverse roles in regulating anxiety, motor function, and sedation. While non-selective GABA_AR positive allosteric modulators (PAMs), such as benzodiazepines, are clinically effective anxiolytic drugs, their non-selective activity across α1/2/3/5 subunit-containing GABA_ARs leads to sedation, cognitive impairment, and risk of dependence. To address this, we evaluated ENX-102, a novel GABA_AR PAM, which exhibits selectivity for α2/3/5 subunits. In rodents, ENX-102 demonstrated dose-dependent anxiolytic-like activity following acute and sub-chronic administration, without sedation. ENX-102 exhibited a dose-dependent quantitative electroencephalography (qEEG) spectral signature in rodents that was distinct from that of benzodiazepines. In a double-blind, placebo-controlled, multiple-ascending dose study in healthy human volunteers, ENX-102 was evaluated using the NeuroCart, a CNS test battery including saccadic peak velocity (SPV), adaptive tracking, pupillometry, body sway, the Bond and Lader Visual Analog Scale (VAS), the Visual Verbal Learning Task (VVLT), and qEEG. ENX-102 produced reductions in SPV that were indicative of central target engagement, with minimal effects on alertness and motor coordination, which is consistent with subtype-selective GABA_AR targeting. Notably, qEEG revealed increased β-band power and decreased δ- and θ-band activity, which were distinct from the spectral profile of non-selective PAMs, supporting translational alignment with preclinical findings. Across dose levels, ENX-102 was well tolerated and exhibited favorable pharmacokinetics. These results support further clinical development of ENX-102 as a next-generation GABA_AR subtype-selective anxiolytic drug. Full article

Bilayer transition metal dichalcogenides (TMDs) are promising materials for next-generation field-effect transistors (FETs) due to their atomically thin structure and favorable transport properties. In this study, we employ density functional theory (DFT) to compute the electronic band structures and phonon dispersions of bilayer WS₂, WSe₂, and MoS₂, and the electron-phonon scattering rates using the EPW (electron-phonon Wannier) method. Carrier transport is then investigated within a semiclassical full-band Monte Carlo framework, explicitly including intrinsic electron-phonon scattering, dielectric screening, scattering with hybrid plasmon–phonon interface excitations (IPPs), and scattering with ionized impurities. Freestanding bilayers exhibit the highest mobilities, with hole mobilities reaching 2300 cm²/V·s in WS₂ and 1300 cm²/V·s in WSe₂. Using hBN as the top gate dielectric preserves or slightly enhances mobility, whereas HfO₂ significantly reduces transport due to stronger IPP and remote phonon scattering. Device-level simulations of double-gate FETs indicate that series resistance strongly limits performance, with optimized WSe₂ pFETs achieving ON currents of 820 A/m, and a 10% enhancement when hBN replaces HfO₂. These results show the direct impact of first-principles electronic structure and scattering physics on device-level transport, underscoring the importance of material properties and the dielectric environment in bilayer TMDs. Full article

To address the issue of sustained deformation in the main roadway surrounding rock triggered by intense movement of overlying strata following the reduction of width of the stopping pillar (WSP) in closely spaced double extra-thick coal seams (CSDECS). Analyze the evolution patterns of abutment pressure, principal stress vector lines, and zones of deviatoric stress concentration (ZDSC) of the main roadways using multi-method approaches. The findings are as follows: As the WSP is reduced, the maximum abutment pressure (MAP) on both sides of the gate roadways’ surrounding rock becomes significantly more asymmetric and intense. The deflection trajectory of the maximum principal stress line (MPSL) in the two coal seams, induced by the advancing underlying panel, follows an approximate inverted ︺ shape. The evolution of the ZDSC and the main roadways in the adjacent working faces all shows three-stage characteristics. For the upper coal seam, it is characterized by crescent-shaped symmetry → slow and asymmetric increase of the peak value and the offset of the ZDSC → the ZDSC on the non-mining side (NM-S) reaches the maximum while the mining side (M-S) shows the reverse trend. For the lower coal seam, it is characterized by crescent-shaped symmetry → quasi-annular distribution with a slight increase in the peak value → significant and asymmetric increase of the peak values. Based on the identification of the key control zones in the ZDSC, an asymmetric reinforcement segmented control method was proposed. The findings provide useful guidance for analogous engineering projects. Full article

Electrolyte-gated transistors (EGTs) have emerged as a highly promising platform for neuromorphic computing and bioelectronics, offering potential solutions to overcome the limitations of the von Neumann architecture. This comprehensive review examines recent advancements in EGT technology, focusing on three critical dimensions: materials, device configurations, and external field regulation strategies. We systematically analyze the development and properties of diverse electrolyte materials, including liquid electrolyte, polymer-based electrolytes, and inorganic solid-state electrolytes, highlighting their influence on ionic conductivity, stability, specific capacitance, and operational characteristics. The fundamental operating mechanisms of EGTs and electric double layer transistors (EDLTs) based on electrostatic modulation and ECTs based on electrochemical doping are elucidated, along with prevalent device configurations. Furthermore, the review explores innovative strategies for regulating EGT performance through external stimuli, including electric fields, optical fields, and strain fields/piezopotentials. These multi-field regulation capabilities position EGTs as ideal candidates for building neuromorphic perception systems and energy-efficient intelligent hardware. Finally, we discuss the current challenges such as material stability, interfacial degradation, switching speed limitations, and integration density. Furthermore, we outline future research directions, emphasizing the need for novel hybrid electrolytes, advanced fabrication techniques, and holistic system-level integration to realize the full potential of EGTs in next-generation computing and bio-interfaced applications. Full article

Discrete Silicon Carbide Metal-Oxide-Semiconductor Field-Effect Transistors (SiC MOSFETs) are characterized by their lower parasitic parameters and single-chip design, enabling them to achieve even faster switching speeds. However, the rapid rate of change in voltage (dv/dt) and current (di/dt) can lead to overshoot and oscillation in both voltage and current, ultimately limiting the performance of high-frequency operations. To address this issue, this paper presents a high-switching-frequency motor controller that utilizes discrete SiC MOSFETs. To achieve a high switching frequency for the controller while minimizing current oscillation and voltage overshoot, a novel electronic system architecture is proposed. Additionally, a passive driving circuit is designed to suppress gate oscillation without the need for additional control circuits. A new printed circuit board (PCB) laminate stack featuring low parasitic inductance, high current conduction capacity, and efficient heat dissipation is also developed using advanced wiring technology and a specialized heat dissipation structure. Compared to traditional methods, the proposed circuit and bus design features a simpler structure, a higher power density, and achieves a 13% reduction in current overshoot, along with a 15.7% decrease in switching loss. The silicon carbide (SiC) controller developed from this research has successfully undergone double-pulse and power testing. The results indicate that the designed controller can operate reliably over extended periods at a switching frequency of 50 kHz, achieving a maximum efficiency of 98.2% and a power density of 9 kW/kg (10 kW/L). The switching frequency and quality density achieved by the controller have not been observed in previous studies. This controller is suitable for use in the development of new energy electrical systems. Full article

The switching characteristics of GaN HEMT devices exhibit a very complex dynamic nonlinear behavior and multi-physics coupling characteristics, and traditional switching current models based on physical mechanisms have significant limitations. This article adopts a hybrid architecture of convolutional neural network and long short-term memory network (CNN-LSTM). In the 1D-CNN layer, the one-dimensional convolutional neural network can automatically learn and extract local transient features of time series data by sliding convolution operations on time series data through its convolution kernel, making these local transient features present a specific form in the local time window. In the double-layer LSTM layer, the neural network model captures the transient characteristics of switch current through the gating mechanism and state transfer. The hybrid architecture of the constructed model has significant advantages in accuracy, with metrics such as root mean square error (RMSE) and mean absolute error (MAE) significantly reduced, compared to traditional switch current models, solving the problem of insufficient accuracy in traditional models. The neural network model has good fitting performance at both room and high temperatures, with an average coefficient close to 1. The new neural network hybrid architecture has short running time and low computational resource consumption, meeting the needs of practical applications. Full article

Background: Peripancreatic collections (PPCs) are a frequent and severe complication of acute and chronic pancreatitis, as well as pancreatic surgery, often requiring interventions to treat and prevent infection, gastric obstruction, and other complications. Endoscopic ultrasound (EUS)-guided drainage has emerged as a minimally invasive alternative to surgical and percutaneous approaches, offering reduced morbidity and shorter recovery times. However, the effectiveness of EUS-guided drainage in post-surgical PPCs remains underexplored. Methods: This retrospective, single-center study evaluated the technical and clinical outcomes of EUS-guided drainage in patients with PPCs between October 2021 and December 2024. Patients were categorized as having post-pancreatitis or post-surgical PPCs. Technical success, clinical success, complications, recurrence rates, and the need for reintervention were assessed. Results: A total of 50 patients underwent EUS-guided drainage, including 42 (84%) with post-pancreatitis PPCs and 8 (16%) with post-surgical PPCs. The overall technical success rate was 100%, with clinical success achieved in 96% of cases. Lumen-apposing metal stents (LAMSs) were used in 84% of patients, including 7.1% as a dual-gate salvage strategy after the failure of double-pigtail drainage. The complication rate was 24%, with infection being the most common (16%). The recurrence rate was 25%, with no significant difference between post-pancreatitis and post-surgical cases. Patients with walled-off necrosis had a significantly higher reintervention rate (35%) than those with pseudocysts (18%; p = 0.042). Conclusions: EUS-guided drainage is a highly effective and safe intervention for PPCs, including complex post-surgical cases. The 100% technical success rate reinforces its reliability, even in anatomically altered post-surgical collections. While recurrence rates remain a consideration, EUS-guided drainage offers a minimally invasive alternative to surgery, with comparable outcomes in both post-pancreatitis and post-surgical patients. Future multi-center studies should focus on optimizing treatment strategies and reducing recurrence in high-risk populations. Full article

Dynamic on-resistance (R_ON) of commercial GaN on Si normally off high-electron-mobility transistor (HEMT) devices is a very important parameter because it is responsible for conduction losses that limit the power conversion efficiency of high-power switching converters. Due to charge trapping effects, dynamic R_ON is always higher than in DC, a behavior known as current collapse. To study how short-time dynamics of charge trapping and release affects R_ON we use rectangular 0–5 V gate voltage pulses with durations in the 1 μs to 100 μs range. Measurements are first carried out for single pulses of increasing duration, and it is found that R_ON depends on both pulse duration and drain current I_D, being higher at shorter pulse durations and lower I_D. For a train of five pulses, R_ON decreases with pulse number, reaching a steady state after a time interval of 100 μs. The response to a five pulses train is compared to that of a square-wave signal to study the time evolution of R_ON toward a dynamic steady state. The DC R_ON is also measured, and it is a factor of ten smaller than dynamic R_ON at the same I_D. This confirms that a reduction in trapped charges takes place in DC as compared to the square-wave switching operation. Additional off-state stress tests at V_DS = 55 V reveal the presence of residual surface traps in the drain access region, leading to four times increase in R_ON in comparison to pristine devices. Finally, the dynamic R_ON is also measured by the double-pulse test (DPT) technique with inductive load, giving a good agreement with results from single-pulse measurements. Full article

The eGaN HEMT power devices face serious crosstalk problems when applied to high-frequency bridge circuits, thereby limiting the switching performance of these devices. To address this issue, a gate driver is proposed in this paper that can suppress both positive and negative crosstalk of eGaN HEMT power devices, offering the advantages of simple control and easy integration. The basic idea is to suppress positive crosstalk by constructing a negative voltage capacitor, and to suppress negative crosstalk by reducing the impedance of the gate loop. To verify the capability of the proposed gate driver, double-pulse and synchronous Buck test platforms are constructed. The experimental results clearly demonstrate that the proposed gate driver reduces the positive and negative crosstalk spikes by 2.03 V and 1.54 V, respectively, ensuring that the positive and negative crosstalk spikes fall within a safe operating range. Additionally, the turn-off speed of the device is enhanced, leading to a reduction in switching loss. Full article

Rain streaks degrade the quality of optical images and seriously affect the effectiveness of subsequent vision-based algorithms. Although the applications of a convolutional neural network (CNN) and self-attention mechanism (SA) in single image deraining have shown great success, there are still unresolved issues regarding the deraining performance and the large computational load. The work in this paper fully coordinates and utilizes the advantages between CNN and SA and proposes a hybrid content-aware deraining network (CAD) to reduce complexity and generate high-quality results. Specifically, we construct the CADBlock, including the content-aware convolution and attention mixer module (CAMM) and the multi-scale double-gated feed-forward module (MDFM). In CAMM, the attention mechanism is used for intricate windows to generate abundant features and simple convolution is used for plain windows to reduce computational costs. In MDFM, multi-scale spatial features are double-gated fused to preserve local detail features and enhance image restoration capabilities. Furthermore, a four-token contextual attention module (FTCA) is introduced to explore the content information among neighbor keys to improve the representation ability. Both qualitative and quantitative validations on synthetic and real-world rain images demonstrate that the proposed CAD can achieve a competitive deraining performance. Full article

In today’s business environment, emails are essential across all sectors, including finance and academia. There are two main types of emails: ham (legitimate) and spam (unsolicited). Spam wastes consumers’ time and resources and poses risks to sensitive data, with volumes doubling daily. Current spam identification methods, such as Blocklist approaches and content-based techniques, have limitations, highlighting the need for more effective solutions. These constraints call for detailed and more accurate approaches, such as machine learning (ML) and deep learning (DL), for realistic detection of new scams. Emphasis has since been placed on the possibility that ML and DL technologies are present in detecting email spam. In this work, we have succeeded in developing a hybrid deep learning model, where Long Short-Term Memory (LSTM) and the Gated Recurrent Unit (GRU) are applied distinctly to identify spam email. Despite the fact that the other models have been applied independently (CNNs, LSTM, GRU, or ensemble machine learning classifier) in previous studies, the given research has provided a contribution to the existing body of literature since it has managed to combine the advantage of LSTM in capturing the long-term dependency and the effectiveness of GRU in terms of computational efficiency. In this hybridization, we have addressed key issues such as the vanishing gradient problem and outrageous resource consumption that are usually encountered in applying standalone deep learning. Moreover, our proposed model is superior regarding the detection accuracy (90%) and AUC (98.99%). Though Transformer-based models are significantly lighter and can be used in real-time applications, they require extensive computation resources. The proposed work presents a substantive and scalable foundation to spam detection that is technically and practically dissimilar to the familiar approaches due to the powerful preprocessing steps, including particular stop-word removal, TF-IDF vectorization, and model testing on large, real-world size dataset (Enron-Spam). Additionally, delays in the feature comparison technique within the model minimize false positives and false negatives. Full article

This study systematically investigates the mechanism by which total ionizing dose (TID) affects the lifetime degradation of NMOS devices induced by hot-carrier injection (HCI). Experiments involved Cobalt-60 (Co-60) gamma-ray irradiation to a cumulative dose of 500 krad (Si), followed by 168 h annealing at 100 °C to simulate long-term stability. However, under HCI stress conditions (V_D = 2.7 V, V_G = 1.8 V), irradiated devices show a 6.93% increase in threshold voltage shift (ΔV_th) compared to non-irradiated counterparts. According to the IEC 62416 standard, the lifetime degradation of irradiated devices induced by HCI stress is only 65% of that of non-irradiated devices. Conversely, when the saturation drain current (I_Dsat) degrades by 10%, the lifetime doubles compared to non-irradiated counterparts. Mechanistic analysis demonstrates that partial neutralization of E’ center positive charges at the gate oxide interface by hot electrons weakens the electric field shielding effect, accelerating ΔV_th drift, while interface trap charges contribute minimally to degradation due to annealing-induced self-healing. The saturation drain current shift degradation primarily correlates with electron mobility variations. This work elucidates the multi-physics mechanisms through which TID impacts device reliability and provides critical insights for radiation-hardened design optimization. Full article