Low-frequency Characteristics Research Articles

Disentangled feature fusion network for lightweight image super-resolution

Recently, the quality of generated images in image super-resolution (SR) has significantly improved due to the widespread application of convolutional neural networks. Existing super-resolution methods often overlook the importance of network width and instead prioritize designing deeper network structures to improve performance. Deeper networks, however, demand greater computational resources and are more challenging to train, making them less suitable for devices with limited performance. To address this issue, we propose a novel method called the disentangled feature fusion network (DFFN). Specifically, we construct dual feature extraction streams (DFES) using the strategies of feature disentanglement and parameter sharing, which increase the network's width without increasing the number of parameters. To facilitate the interaction of information between the dual feature extraction streams, we design a feature interaction module (FIM) that separately enhances high and low-frequency features. Furthermore, a feature fusion module (FFM) is presented to efficiently fuse different levels of high and low-frequency feature information. The proposed DFFN integrates DFES, FIM, and FFM, which not only widens the network without increasing parameters but also minimizes the loss of crucial feature information. Extensive experiments on five benchmark datasets demonstrate that our method significantly outperforms other state-of-the-art lightweight super-resolution methods, achieving a superior balance between parameter quantity, reconstruction performance, and model complexity. The code is available at https://github.com/zhoujy-aust/DFFN.

Prediction of Transonic Shock Buffet over Supercritical Airfoil OAT15A Based on Zonal Detached-Eddy Simulation

The zonal detached-eddy simulation (ZDES) method divides the flow field into different computational modes, offering flexibility in the simulation of complex flows. The transonic shock buffet on the upper surface of the OAT15A supercritical airfoil is numerically simulated by ZDES based on the k-ω SST model. The results show that the error in the low-frequency characteristics of the transonic shock buffet predicted by ZDES compared to experiments is within 3%. Its predictions of pressure fluctuations and averaged velocity agree better with the experimental data. Overall, ZDES is effective in predicting the periodic oscillations of shock waves along the streamwise direction, flow separation induced by shock wave motion, and the small-scale vortex structures in the wake region.

Motor-Related EEG Analysis Using a Pole Tracking Approach.

This study introduces an alternative approach to electroencephalography (EEG) time-frequency analysis based on time-varying autoregressive (TV-AR) models in a cascade configuration to independently monitor key EEG spectral components. The method is evaluated for its neurophysiological interpretation and effectiveness in motor-related brain-computer interface (BCI) applications. Specifically, we assess the ability of the tracked EEG poles to discriminate between rest, movement execution (ME) and movement imagination (MI) in healthy subjects, as well as movement attempts (MA) in individuals with spinal cord injury (SCI). Our results show that pole tracking effectively captures broad changes in EEG dynamics, such as transitions between rest and movement-related states. It outperformed traditional EEG-based features, increasing detection accuracy for ME by an average of 4.1% (with individual improvements reaching as high as 15%) and MI by an average of 4.5% (up to 13.8%) compared to time-domain low-frequency EEG features. Similarly, compared to alpha/beta band power, the method improved ME detection by an average of 5.9% (up to 10.4%) and MI by an average of 4.3% (up to 10.2%), with results averaged across 15 healthy participants. In one participant with SCI, pole tracking improved MA detection by 12.9% over low-frequency EEG features and 4.8% over alpha/beta band power. However, its ability to distinguish finer movement details within specific movement types was limited. Additionally, the temporal evolution of the extracted pole tracking features revealed event-related desynchronization phenomena, typically observed during ME, MA and MI, as well as increases in frequency, which are of neurophysiological interest.

Reconfigurable topological gradient metamaterials and potential applications

Elastic topological metamaterials are innovative fields of the fusion of physics and mechanics. Through novel microstructural designs, elastic topological metamaterials provide a pioneering approach to precisely manipulate elastic waves, maintaining robustness and efficiency in wave propagation even within complex environments. Recent advancements in reconfigurability and gradient features have significantly enriched the mechanical phenomena of wave manipulation, broadening the application potential of elastic topological metamaterials. In this study, we design the local resonant phononic crystal, employing a spiral support structure to ensure low-frequency characteristics and adjustable mass for flexible configuration. By utilizing the designed lattices with distinct topological phases, the topological valley edge states are constructed. The frequency and group velocity variation of the topological edge states under different stiffness and mass parameters are analyzed. Furthermore, we further constructed the mass and stiffness gradient metamaterial to analyze the transmission law of elastic waves in topological metamaterials. Additionally, the fluctuation features of the wave transmittance under disordered and impurity defects were discussed, confirming the robustness of waveguides within low-frequency topological metamaterials. Finally, we explored the potential applications of the designed metamaterials in energy localization and low-frequency reconfigurable waveguides. Numerical analysis showed that the topological gradient metamaterials can enhance the vibration energy at a specific interface, and low-frequency bend waveguide paths can be adjusted flexibly by configurable mass. In summary, this paper focuses on the low-frequency and gradient features of elastic topological metamaterials, aiming to unlock their application potential in vibrational energy harvesting, vibration suppression, and information transmission, which accelerate the practical implementation of topological metamaterials.

Full-scale test of disc cutter rotary cutting in TBM tunnelling: a case study of Mawan granite in Shenzhen, China

Understanding the distribution of cutting force provides a foundation for the design and optimization of cutterhead layout. The simulation of rotary cutting with a large cutting radius remains a challenge for laboratory tests. In this study, taking the driving condition of the Mawan Tunnel located in the Guangdong-Hong Kong-Macao Greater Bay Area as a case study, a full-scale rotary cutting platform with large cutting radius was used to investigate the rock-breaking effects. Rotary cutting with cutting depths of 1–8 mm and cutting radii of 200–1400 mm were carried out on the granite testing platform using a 19-in disc cutter. The normal force, rolling force, and side force during the cutting process were systematically monitored. Moreover, the time–frequency characteristics of the rock breaking process and the effects of cutting depth, cutting radius and stress state on the average and peak loads were investigated. The results show that the average normal force and rolling force increased with increasing cutting depth, while the side force was less affected by cutting depth through the Boltzman model analysis. The average normal force and rolling force were not significantly affected by the cutting radius, but the average side force decreased as a power function with increasing cutting radius, and showed a stable side force when the installation radius exceeded 1000 mm. The dominant frequency of the cutting force was 0–1.2 Hz, exhibiting notable low-frequency characteristics. Additionally, the ratio of peak to average (RPA) with respect to triaxial cutting force demonstrated an increase with the growth of cutting depth, while RPA in terms of side force exhibited a corresponding growth with the increasing cutting radius. In particular, the cutting force characteristics of the cutterhead obtained in this study can inform the layout of disc cutters and real-life TBM operations, thereby avoiding severe overloading of internal disc cutters.

LapUNet: a novel approach to monocular depth estimation using dynamic laplacian residual U-shape networks

Monocular depth estimation is an important but challenging task. Although the performance has been improved by adopting various encoder-decoder architectures, the estimated depth maps lack structure details and clear edges due to simple repeated upsampling. To solve this problem, this paper presents the novel LapUNet (Laplacian U-shape networks), in which the encoder adopts ResNeXt101, and the decoder is constructed with the novel DLRU (dynamic Laplacian residual U-shape) module. The DLRU module based on the U-shape structure can supplement high-frequency features by fusing dynamic Laplacian residual into the process of upsampling, and the residual is dynamically learnable due to the addition of convolutional operation. Also, the ASPP (atrous spatial pyramid pooling) module is introduced to capture image context at multiple scales though multiple parallel atrous convolutional layers, and the depth map fusion module is used for combining high and low frequency features from depth maps with different spatial resolution. Experiments demonstrate that the proposed model with moderate model size is superior to other previous competitors on the KITTI and NYU Depth V2 datasets. Furthermore, 3D reconstruction and target ranging by utilizing the estimated depth maps prove the effectiveness of our proposed method.

Repairing Noise-Contaminated Low-Frequency Vibrational Spectra with an Attention U-Net.

Low-frequency vibrational modes in infrared (IR) and Raman spectra, often termed molecular fingerprints, are sensitive probes of subtle structural changes and chemical interactions. However, their inherent weakness and susceptibility to environmental interference make them challenging to detect and analyze. To tackle this issue, we developed a deep learning denoising protocol based on an attention-enhanced U-net architecture. This model leverages the inherent correlations between high- and low-frequency vibrational modes within a molecule, effectively reconstructing low-frequency spectral features from their high-frequency counterparts. We demonstrate the effectiveness of this method by recovering low-frequency signals of trans-1,2-bis(4-pyridyl)ethylene (BPE) adsorbed on an Ag surface, a representative system for surface enhancement Raman spectroscopy (SERS). Notably, the trained model exhibits promising transferability to SERS spectra acquired under different surface and external field conditions. Furthermore, we applied this method to experimental IR and Raman spectra of BPE, achieving high-quality, low-frequency spectral recovery.

Interictal stereo-electroencephalography features below 45 Hz are sufficient for correct localization of the epileptogenic zone and postsurgical outcome prediction.

Evidence suggests that the most promising results in interictal localization of the epileptogenic zone (EZ) are achieved by a combination of multiple stereo-electroencephalography (SEEG) biomarkers in machine learning models. These biomarkers usually include SEEG features calculated in standard frequency bands, but also high-frequency (HF) bands. Unfortunately, HF features require extra effort to record, store, and process. Here we investigate the added value of these HF features for EZ localization and postsurgical outcome prediction. In 50 patients we analyzed 30 min of SEEG recorded during non-rapid eye movement sleep and tested a logistic regression model with three different sets of features. The first model used broadband features (1-500 Hz); the second model used low-frequency features up to 45 Hz; and the third model used HF features above 65 Hz. The EZ localization by each model was evaluated by various metrics including the area under the precision-recall curve (AUPRC) and the positive predictive value (PPV). The differences between the models were tested by the Wilcoxon signed-rank tests and Cliff's Delta effect size. The differences in outcome predictions based on PPV values were further tested by the McNemar test. The AUPRC score of the random chance classifier was .098. The models (broad-band, low-frequency, high-frequency) achieved median AUPRCs of .608, .582, and .522, respectively, and correctly predicted outcomes in 38, 38, and 33 patients. There were no statistically significant differences in AUPRC or any other metric between the three models. Adding HF features to the model did not have any additional contribution. Low-frequency features are sufficient for correct localization of the EZ and outcome prediction with no additional value when considering HF features. This finding allows significant simplification of the feature calculation process and opens the possibility of using these models in SEEG recordings with lower sampling rates, as commonly performed in clinical routines.

Deep sediment heterogeneity inferred using very low-frequency features from merchant shipsa).

The very low-frequency noise from merchant ships provides a good broadband sound source to study the deep layers of the seabed. The nested striations that characterize ship time-frequency spectrograms contain unique acoustic features corresponding to where the waveguide invariant β becomes infinite. In this dataset, these features occur at frequencies between 20 and 80 Hz, where pairs of modal group velocities become equal. The goal of this study is to identify these β = ∞ frequencies in ship noise spectrograms and use them to perform statistical inference for the deep layer sound speeds and thicknesses in the New England Mudpatch for a larger number of ships and acoustic arrays over a larger geographical region than previously studied. Marginal probability distributions of the data indicate that using singular points for a feature-based inversion yields an estimate of the sound speed and a limiting value for the thickness of the first deep layer. Heterogeneity is examined by correlating spatial variability of the deep layer sound speeds with ship tracks.

Prediction of Pressure Fluctuation of Double-Suction Centrifugal Pump in Different Cavitation Conditions based on EMD-GRU

Abstract The double-suction centrifugal pumps play a crucial role in various applications such as in irrigation-drainage projects, water diversion projects, and urban water supply and drainage projects. The pressure fluctuation is one of the crucial parameters for assessing the operational stability of pump. The pressure fluctuation exhibits distinct characteristics in various cavitation conditions. A pressure fluctuation prediction method based on Empirical Mode Decomposition (EMD) and Gated Recurrent Unit (GRU) is proposed in this paper. The model enables accurate forecasting of pressure fluctuation signals when pump operates at different cavitation conditions. The data were obtained from experimental measurements of a double-suction centrifugal pump. The original signal was decomposed into multiple sub-signals with varying energy levels using the EMD method. The neural network of GRU was employed for training and data prediction on individual sub-signals. The predicted results of each sub-signal were combined to form the final pressure fluctuation signals. The method based on EMD-GRU proves effective in predicting pressure fluctuation in three different cavitation conditions. The mean absolute error (MAE) in time-domain characteristics is below 6%, and the prediction error for the main frequency is below 1%. The prediction of low-frequency characteristics still requires improvement.

AMTrack:Transformer tracking via action information and mix-frequency features

Nowadays, Transformer-based visual tracking algorithms have been developing quickly because of the self-attention mechanism of Transformer, which has the capability to model global information. Although the self-attention mechanism in Transformer can effectively capture long-range dependencies in feature space, they only use flattened two-dimensional features and are unable to capture long-range temporal dependencies. Furthermore, since the self-attention in Transformer functions as a low-pass filter, it picks up on low-frequency features of the target while ignoring high-frequency features. This research suggests a Transformer tracker based on action information and mix-frequency features (AMTrack) to address these problems. Specifically, to address the lack of temporal remote dependencies, we introduce the target action aware module and the target action offset module. The target action aware module sets up several pathways to extract spatio-temporal, channel, and motion feature independently. In contrast, the target action offset module computes the target’s offset information by computing relative feature maps. Furthermore, in order to address the imbalance between high and low frequency features, we propose a mix-frequency attention and multi-frequency self-attention convolutional block. The mix-frequency attention uses high-frequency features within partitioned local windows as input for the high-frequency branch and average-pooled low-frequency features as the input for the low-frequency branch, calculating attention scores in both branches respectively. The multi-frequency self-attention convolutional block uses self-attention to capture low-frequency features and convolution to capture high-frequency features. Extensive experiments are carried out on eight challenging tracking datasets (e.g., OTB100 (Object Tracking Benchmark 100), NFS (Need For Speed), UAV123 (Unmanned Aerial Vehicles 123), TC128 (Temple Color 128), VOT2018 (Visual Object Tracking 2018), LaSOT (Large-scale Single Object Tracking), TrackingNet (Tracking Network), GOT-10k (Generic Object Tracking-10k)), and the experimental results show that our tracker achieves excellent tracking performance when compared with several state-of-the-art tracking algorithms. The experimental results show that on LaSOT, the success rates AUC (Area Under Curve), PNorm, and P reach 65.8%, 69.2%, and 68.0%, respectively, where the AUC value is 2.1% higher than the baseline algorithm TrDiMP (Transformer Discriminative Model Prediction). On other datasets, our tracker also achieves excellent tracking performance.

Diversity and transition of periodic motion of a periodically excited soft-impacting machinery

Dynamics of a periodically excited vibro-impact system with soft impacts is investigated. Essential features of period-one multi-impact motion group and correlated transition characteristics in low-frequency range are discussed in detail by the way of two-parameter bifurcation space providing qualitative domains for different periodic motions. The main focus is given to the effect of sensitive parameters including constraint stiffness k 0, clearance threshold b, and damping parameter ζ on the system response. The low-frequency characteristics in the finite-dimensional parameter space are particularly explored. It is found that the increase of k 0 induces multi-type bifurcation of period-one double-impact symmetrical motion, which induces a rich variety of periodic motions, and period-one multi-impact motion group orbit primarily exist in the small-clearance b and low-frequency ω zone. Based on the evolution irreversibility of adjacent period-one multi-impact orbit, the mechanism of singularies appearing in pairs and two different transition zones (hysteresis and liguliform zones) is studied, the result of which provides a theoretical reference value for the common low-frequency vibration instability phenomenon in the field of mechanical engineering. For small-damping coefficient ζ, period-one multi-impact motion has a large quantity, and the main bridge for the transition of adjacent period-one multi-impact motion is liguliform zone, which embraces period-one multi-impact asymmetrical motion and period-n multi-impact subharmonic motion and a certain chaotic zone. For large-damping coefficient ζ, the amount of period-one multi-impact motion group is reduced, and the main bridge for the transition of adjacent period-one multi-impact motion is hysteresis zone, where adjacent period-one multi-impact orbits can coexist according to initial conditions. As designing and renovating impact mechanical equipment, the reasonable matching law of dynamic parameters can be determined through two-parameter bifurcation space, which is conducive to making the system work in stable periodic motion and obtaining larger instantaneous impact velocity.

Sodium-Poor, Hydroxyl-Rich, Defective Na2Ti3O7 Prepared by γ-Irradiation and Its Enhanced Proton Conductivity.

The use of γ-irradiation to tailor the physicochemical properties of materials is not widely applied to layered alkali metal oxides. Herein, we show that γ-irradiation (up to 400 kGy) of Na2Ti3O7 leads to a sodium-poor, hydroxyl-rich analogue where the layered structure, plate-like morphology, and textural properties are preserved. The deintercalation of sodium ions modifies the Ti-O bond lengths and expands the unit cell; the latter is supported by density functional theory (DFT) calculations. 23Na solid-state NMR suggests the transport of the symmetric, 7-fold Na2 sites to an intermediate environment, which is closer to the asymmetric, 9-fold Na1 sites. An 8 wt % mass loss (1.4 mol water/mol titanate) is observed, indicating an increased concentration of protons/hydroxyls. These hydroxyl groups (i.e., lattice protons) possess higher thermal stability than solely surface-adsorbed ones in the nonirradiated sample. At 200-400 kGy, the proton conduction (50 °C and ∼70% RH) of ∼10-6 S·cm-1 is 1 order of magnitude larger than that in the nonirradiated sample; the relaxation time decreases from 30 to 2-6 μs with γ-irradiation. The γ-dose dependence of dielectric loss is also present and analyzed using the Jonscher universal power law, indicating the low-frequency dispersion behavior characteristics of high charge densities.

CTAFFNet: CNN–Transformer Adaptive Feature Fusion Object Detection Algorithm for Complex Traffic Scenarios

As the core technology of an environmental perception system, object detection has received more and more attention and has become a hot research direction for intelligent driving vehicles. The CNN–Transformer hybrid model has poor generalization ability, making it difficult to meet the detection requirements for small objects in complex scenes. We propose a novel convolutional neural network (CNN)–Transformer Adaptive Feature Fusion Network (CTAFFNet) for object detection. First, we design a Local–Global Feature Fusion unit known as the Convolutional Transformation Adaptive Fusion Kernel (CTAFFK), which is integrated into CTAFFNet. The CTAFFK kernel utilizes two branches, namely CNN and Transformer, to extract local and global features from the image, and adaptively fuses the features from both branches. Subsequently, we develop an adaptive feature fusion strategy that combines local high-frequency and global low-frequency features to obtain comprehensive feature information. Finally, CTAFFNet employs an encoder–decoder structure to facilitate the flow of fused local–global information between different stages, ensuring the model’s generalization capabilities. Results from the experiment conducted on the large and challenging KITTI dataset demonstrate the effectiveness and efficiency of the proposed network. Compared with other mainstream networks, it achieves an average precision of 91.17%, particularly excelling in the detection of small objects at longer distances with a remarkable 70.18% accuracy, while also providing real-time detection capabilities.

A one-step fabrication method of flexible magnetostrictive fiber ribbon based on Fe50Co50 alloy and its characteristics study

Magnetostrictive materials are widely used in sensors, actuators and micro-nano electronics because of their excellent performance. However, traditional rigid magnetostrictive materials are difficult to adapt to the flexible and deformable structure of flexible electronic devices. In order to solve the deformation, sensing and control problems in the field of flexible electronics, a one-step new method of is proposed to deposit Fe50Co50 fiber ribbons with reciprocated and loop structures on flexible substrate. Helmholtz coil is designed to test the magnetostrictive positive effect ability of Fe50Co50 ribbon, and its magnetostrictive strain curve is analyzed. Furthermore, frequency domain characteristics of Fe50Co50 ribbon are tested, and the influence of externally driving magnetic field on elastic layer materials and resonance output characteristics is analyzed. The experimental results show that Fe50Co50 ribbon exhibits good low-frequency magnetic field characteristics. And free end displacement of Fe50Co50 ribbon based on loop structure, under 12 mT alternating magnetic field and 134 mT bias magnetic field, reaches a peak-to-peak value of 870 μm for first-order resonance, and 320 μm for second-order resonance. This verifies that Fe50Co50 ribbon possesses the expected properties that magnetostrictive materials should have, and innovatively breaks through the technical bottleneck of manufacturing flexible micro-nano magnetostrictive energy exchange components.

Low-frequency broadband acoustic metamaterial absorber based on nested resonator and synergistic coupled weak resonances

Utilizing absorbers with sub-wavelength thickness for low-frequency broadband noise control is a challenging problem, and rapid on-demand design of structures based on target noise frequency has received continuous attention. In this work, nested cavities are introduced into the typical neck-embedded Helmholtz resonator (NEHR) to obtain the nested neck-embedded Helmholtz resonator (NNEHR), which exhibits better cavity partitioning volume fraction and covers wider sound absorption band. Both theoretical analysis and simulations demonstrate that NNEHR possesses superior acoustic properties: generating an additional absorption peak, lowering the position of the first absorption peak by 11 %, and increasing the absorption coefficient by 4.7 %. Instantaneous velocity fields and thermal-viscous energy dissipation density illustrate that the efficient sound absorption of NNEHR originates from resonant dissipation in the neck and surrounding areas. Furthermore, by combining the theoretical model with genetic algorithms, three sets of NNEHR metasurfaces are optimized. The optimized structures achieve average absorption coefficients of 0.831, 0.913, and 0.885 in the range of 250–550 Hz, 400–1000 Hz, and 500–2000 Hz, with a thickness of only 50 mm, reflecting sub-wavelength, low-frequency broadband sound absorption characteristics. The excellent absorption stems from synergistic coupled weak resonances and higher-order absorption peaks of constituent units, and the overdamping characteristic ensures the realization of the optimal structure with minimal thickness. The contribution of weak resonant units to the total sound absorption performance can be measured by the surface admittance normalized by the total incident area, and the essence of the coupling effect is the strong interaction between units with adjacent absorbing frequencies. The proposed structure exhibits good sound absorption performance and great adjustability, while the genetic algorithm optimization based on the theoretical model demonstrates high efficiency and robustness, providing a methodology for the design and on-demand optimization of low-frequency broadband structures.

Industrial image anomaly detection based on multi Gaussian discriminant model and robust core set

Abstract To address the issue of false positive (FP) detections in image anomaly detection caused by the loss of low-frequency features when dealing with high-dimensional feature distributions, we propose the multi-layer Gaussian discriminant anomaly detection model (MGAD). This model utilizes distance metrics based on multiple normal distributions to perform anomaly detection. By mining multi-layer feature combinations from normal samples and incorporating a Gaussian mixture model strategy for pixel-by-pixel probability density estimation, a weighting mechanism is designed to emphasize the role of low-frequency features in Gaussian space. This approach effectively models data collections that do not follow a single normal distribution as a mixture of several Gaussian distributions, thereby reducing false detections. Additionally, we propose a method for calculating the minimum Mahalanobis distance based on the estimation of the minimum covariance determinant. By identifying a subset with the smallest covariance matrix determinant, this method enhances the robust estimation of the data’s central position and spread, thereby reducing the impact of outliers. On the MVTec-AD dataset, MGAD demonstrates outstanding performance with an anomaly detection area under the receiver operating characteristic curve (AUROC) of 98.8%, the anomaly localization AUROC of 98.2%, and the per-class true negative rate for normal samples of 93.1%. Compared with the state-of-the-art models, MGAD improves the detection accuracy for normal samples by 3.6%, demonstrating the best performance among all models. These results highlight the model’s excellent capability in anomaly recognition and reduction of FPs.

Impact of Oxide Charges on The Minority Carrier Response in MOS(p) Devices With Al2O3/SiO2 Gate Stacks under Strong Inversion Condition

Admittance measurement, including capacitance and conductance, is very useful to analyze the electrical characteristics of metal-oxide-semiconductor (MOS) systems. These techniques offer insights into various device properties, including interface trap density and minority carrier lifetime in the substrate. It is essential to note that, apart from the gate area, the oxide region beyond the device could significantly influence measurement results [1]-[2]. In this work, a compact model proposed in [3] was used to clarify an unusual minority carrier response time observed in experimental data. The results proved the heightened sensitivity of device alternating current (AC) characteristics to oxide charges in the outer region. Additionally, the dependencies of oxide thickness and quality on the transition frequency (ωm) of devices were explored.The MOS (Al/AlOx/SiO2/p-Si) structure illustrated in Fig. 1 was employed in this study. Aluminum oxide region is under the gate electrode, while silicon oxide (SiO2) covers the entire wafer. Table 1 summarizes samples (S1 to S5) with consistent aluminum oxide thickness but diverse SiO2 thickness. Fig 2. shows the process flow and the anodic oxidation (ANO) system. By tilting the wafer at a specific angle during ANO, a high-quality SiO2 layer with different thicknesses on a single wafer could be obtained. Fig 3(a) to (d) illustrate the capacitance-voltage (C-V) and conductance-voltage (G-V) characteristics of devices with varying SiO2 thicknesses. All devices exhibit low-frequency capacitance characteristics in strong inversion regions with thickness-dependent minority carrier response, particularly at 10kHz. Additionally, devices show different thicknesses depending on G-V characteristics at 1MHz and 10kHz. Fig 4(a) and (b) illustrate the Cinv/Cox ratio for varied oxide thicknesses and Gm/ω characteristics under strong inversion bias (VG = +3V). Gm means the measurement conductance. Following established research [3], peak Gm/ω magnitudes would occur at ωm. Table 2 reveals microsecond-range response times (τR) at ωm, considerably faster than conventional silicon minority carrier lifetimes (usually 1 ms to 100 ms). Moreover, decreasing oxide thickness shifts the ωm toward lower values.To explain this observed phenomenon, the environment outside the devices should be taken into consideration. Fig 5(a) and (b) show the schematic diagram and AC signal model proposed in [1] and [2]. Oxide charges in the outer region would deplete majority carriers and invert the p-type silicon surface to an n-type inversion layer. As a result, the lateral depletion region (Cd, out) and conductive channel (Rs) are formed outside the device, which are functions of substrate doping concentration and the effective oxide charges. The detailed model derivation is discussed in [2]. For this reason, an AC signal applied on the gate would extend outside the device for about a millimeter to a micrometer, affecting measuring results. Based on the model, the effective charges in the SiO2 could be extracted by measuring results. The thicker oxide has more oxide charges, as shown in Fig 6(a). Fig 6(b) and (c) show the modeling results of the Cinv/Cox ratio versus different oxide thicknesses and the Gm/ω-ω plot, which are very close to the experimental results. By simplifying the conductance formula (Gm), Gm reveals an inverse proportional to lateral effective conductance (GLeff) at low frequency (e.g., 10 kHz) and a direct proportional at high frequency (e.g., 1 MHz). Thicker SiO2 devices possess more oxide charges, leading to a greater lateral coupling effect, resulting in a greater GLeff compared to thinner devices. This elucidates the observed variations in Gm-V characteristics at 1 MHz and 10 kHz. Figures 7(a) and (b) display TCAD-simulated C-V and ωm for devices with varying oxide charge densities, confirming that minor changes in oxide charges result in noticeable differences in minority carrier responses in the strong inversion region. Fig 8(a) and (b) show the TCAD simulated results for devices with low doping substrate. For low doping substrates, oxide charges play a much more important role under AC operation.In summary, the study established a correlation between oxide charge densities and thicknesses. The origin of the faster minority carrier response and the peak shift in Gm/ω- ω plot were thoroughly explained and validated by TCAD simulation. Oxide charges were found to play a crucial role, especially in low-doping substrates, offering insights for advanced IC fabrication.[1] E. H. Nicollian and A. Goetzberger, IEEE Transactions on Electron Devices, vol. 12, no. 3, pp. 108-117, 1965.[2] K. C. Chen, K. W. Lin, S. W. Huang, J. Y. Lin, and J. G. Hwu, IEEE Journal of the Electron Devices Society, vol. 10, pp. 960-969, 2022.[3] S. Monaghan et al., IEEE Transactions on Electron Devices, vol. 61, no. 12, pp. 4176-4185, 2014. Figure 1

Geometric-aware RGB-D representation learning for hand–object reconstruction

Reconstructing hand–object interaction from single images, crucial for interactive applications, is challenging due to the diversity of hand–object poses and shapes. Depth maps effectively complement RGB data in understanding these interactions geometrically within challenging scenes. However, most existing methods do not fully utilize the potential benefits of RGB-D information fused at different feature levels for reconstruction, limiting their capability to capture the geometric details of hand–object interaction. To address this, we propose an implicit geometric-aware RGB-D representation learning approach using adaptive bidirectional RGB-D feature fusion (ABF) and geometric Fourier feature encoding (GFFE). This method innovatively leverages the hierarchical and complementary high-level features of RGB-D information to enhance the neural implicit representations of hand–object interaction during the feature extraction and encoding stages of RGB-D data. Initially, a two-stream RGB-point cloud encoder, combining CNN and Transformer architectures, extracts appearance and geometric information from RGB and point cloud data. Subsequently, ABF fuses this information, generating dense RGB-D features by leveraging the complementary of appearance and geometric features along with the varying sensitivities of different layers. Finally, GFFE introduces frequency domain information to capture both high-frequency and low-frequency features of objects equally to precisely capture the geometric details of hand–object interaction. Experiments conducted on the real-world DexYCB and the synthetic ObMan benchmarks demonstrate that our approach significantly outperforms existing approaches.

Low-frequency Noise Reduction Mechanism and Acoustic Characteristics of the Dipole Surface Source Structure

Low-frequency Noise Reduction Mechanism and Acoustic Characteristics of the Dipole Surface Source Structure