Suppression Method Research Articles

Score filtering for contextualized noise suppression of Poisson‐distributed geophysical signals

AbstractGeophysical and remote sensing products that rely on Poisson‐distributed measurement signals, such as cosmic‐ray neutron sensing (CRNS) and gamma spectrometry, often face challenges due to inherent Poisson noise. Common techniques to enhance signal stability include data aggregation or smoothing (e.g., moving averages and interpolation). However, these methods typically reduce the ability to resolve detailed temporal (stationary data) and spatial (mobile data) features. In this study, we introduced a method for contextual noise suppression tailored to Poisson‐distributed data, utilizing a discrete score attribution system. This score filter evaluates each observation against eight different criteria to assess its consistency with surrounding values, assigning a score between 0 (very unlikely) and 8 (very likely) to indicate whether the observation is likely to act as noise. These scores can then be used to flag or remove data points based on user‐defined thresholds. We tested the score filter's effectiveness on both stationary and mobile CRNS data, as well as on gamma‐ray spectrometry and electromagnetic induction (EMI) recordings. In our examples, the score filter consistently outperformed established filters, for example Savitzky–Golay and Kalman, in direct competition when applied to CRNS time series data. Additionally, the score filter substantially reduced Poisson noise in mobile CRNS, gamma‐ray spectrometry and EMI data. The scoring system also provides a context‐sensitive evaluation of individual observations or aggregates, assessing their conformity within the dataset. Given its general applicability, customizable criteria and very low computational demands, the proposed filter is easy to implement and holds promise as a valuable tool for denoising geophysical data and applications in other fields.

Research on Radial Vibration Model and Low-Frequency Vibration Suppression Method in PMSM by Injecting Multiple Symmetric Harmonic Currents

Driven by frequency conversion, the windings of a three-phase permanent magnet synchronous motor (PMSM) contain both odd and even harmonic currents. Due to the motor’s pole–slot conductance modulation, the interaction between the magnetic fields generated by these harmonic currents and the permanent magnet field results in harmonic radial vibrations of the motor. This paper analyzes the three-phase currents of the prototype and derives the radial magnetomotive force (MMF) spatiotemporal models for symmetric harmonic currents. By integrating Maxwell’s magnetic force formula and vibration response formula, the radial vibration models for symmetric harmonic currents are developed. The characteristics of vibrations caused by odd and even harmonic currents, as well as positive sequence and negative sequence harmonic currents, are analyzed separately. A cyclic sequence, low-frequency vibration suppression control method incorporating multiple harmonic current injections was designed. Experimental results of this method are compared with those obtained using an ideal sinusoidal current. Except for the second harmonic vibration, all other vibrations are significantly suppressed, with a maximum suppression rate of 92.28%. The total vibration level is reduced by 12.7619 dB, and the average torque is reduced by 0.67% with the total harmonic distortion of the current at 2.89%. The experimental results show that the vibration method in this paper has little influence on the average torque of the motor, the current distortion rate is small, and the vibration suppression effect is good.

Benchmarking Digital Quantum Simulations Above Hundreds of Qubits Using Quantum Critical Dynamics

The real-time simulation of large many-body quantum systems is a formidable task, that may only be achievable with a genuine quantum computational platform. Currently, quantum hardware with a number of qubits sufficient to make classical emulation challenging is available. This condition is necessary for the pursuit of a so-called quantum advantage, but it also makes verifying the results very difficult. In this paper, we flip the perspective and utilize known theoretical results about many-body quantum critical dynamics to qualitatively benchmark digital quantum hardware and various error mitigation techniques. In particular, we benchmark against known universal scaling laws in the Hamiltonian simulation of a time-dependent transverse-field Ising Hamiltonian of up to 133 qubits. Incorporating only basic error mitigation and suppression methods, our study shows reliable control up to a two-qubit gate depth of 28, featuring a maximum of 1396 two-qubit gates, before noise becomes prevalent. These results are transferable to applications such as Hamiltonian simulation, variational algorithms, optimization, or quantum machine learning. We demonstrate this on the example of digitized quantum annealing for optimization and identify an optimal working point in terms of both circuit depth and time step on a 133-site optimization problem. Published by the American Physical Society 2024

Development of New Spray Dust Suppression Materials in Metal Mines and Prediction of Algorithm Simulation Effect

PROBLEM: Dust contamination in metal mining poses substantial dangers to environmental quality and human health. Modern mining operations cannot use traditional spray dust suppression methods because they are poorly adapted to changing climate conditions, low efficient, and detrimental to the environment. INTRODUCTION: Dust pollution seriously impacts the environment and human health in metal mine operations. Traditional spray dust suppression technology has many problems, such as limited effect, environmental impact, and poor climate adaptability. OBJECTIVES: The purpose of this article is to develop a new type of spray dust suppression material and predict its dust suppression effect through algorithm simulation. Firstly, efficient and environmentally friendly dust-reducing materials were screened, and after evaluating the dust-reducing effect under laboratory conditions, the optimal material combination was determined. METHODS: Using computational fluid dynamics (CFD), a numerical model of the spray process was constructed to simulate the dust suppression effect of different materials under different climatic conditions. RESULTS: The results show that the highest dust reduction efficiency of the new spray dust reduction material is more than 4.3% higher than that of the traditional material, and it shows good stability. CONCLUSION: The new spray dust control material and its effect prediction method studied in this article provide an effective solution for dust control in metal mines, which has important theoretical value and practical application prospects.

Π-shaped pixel pattern for shift multiplexing surface holograms in holographic data storage

In this study, a method for shift multiplexing surface holograms was proposed for data storage systems. Shift multiplexing of surface holograms is inhibited by noise specific to Raman–Nath diffraction. We proposed a model to analyze the influence of noise. Subsequently, we proposed a method for noise suppression to arrange signal and reference pixels in a shape similar to the Greek letter “Π.” The proposed method is generally applicable to thin holograms in which Raman–Nath diffraction occurs. The effectiveness of noise suppression using the proposed method was experimentally demonstrated. We theoretically estimated that a shift multiplexing surface holographic data storage system using the proposed method could achieve a significantly higher data transfer rate than conventional optical discs while maintaining a comparable storage density.

Capturing forceful interaction with deformable objects using a deep learning-powered stretchable tactile array

Capturing forceful interaction with deformable objects during manipulation benefits applications like virtual reality, telemedicine, and robotics. Replicating full hand-object states with complete geometry is challenging because of the occluded object deformations. Here, we report a visual-tactile recording and tracking system for manipulation featuring a stretchable tactile glove with 1152 force-sensing channels and a visual-tactile joint learning framework to estimate dynamic hand-object states during manipulation. To overcome the strain interference caused by contact with deformable objects, an active suppression method based on symmetric response detection and adaptive calibration is proposed and achieves 97.6% accuracy in force measurement, contributing to an improvement of 45.3%. The learning framework processes the visual-tactile sequence and reconstructs hand-object states. We experiment on 24 objects from 6 categories including both deformable and rigid ones with an average reconstruction error of 1.8 cm for all sequences, demonstrating a universal ability to replicate human knowledge in manipulating objects with varying degrees of deformability.

High-speed in vivo calcium recording using structured illumination with self-supervised denoising

High-speed widefield fluorescence imaging of neural activity in vivo is fundamentally limited by fluctuations in recorded signal due to background contamination and stochastic noise. In this study, we show background and shot noise-reduced imaging of the ultrafast genetically encoded Ca2+ indicator GCaMP8f in CA1 pyramidal neurons using periodic structured illumination (SI) with computational image reconstruction. We implement what we believe to be a novel reconstruction method for data acquired using periodic structured illumination, termed pseudo-HiLo (pHiLo), that combines a pseudo-widefield (pWF) reconstruction with individual SI frames to perform a HiLo reconstruction. We compare this new technique to interleaved optical sectioning structured illumination microscopy (OS-SIM) and pWF reconstruction. We quantify the performance of each reconstruction by evaluating contrast, transient peak-to-noise ratio (PNR), pairwise correlation coefficients between ΔF/F time courses extracted from individual in-focus cells, and correlation coefficients between each cell with surrounding cell-free background pixels. We additionally incorporate a self-supervised deep learning method for real-time noise suppression (DeepCAD-RT) into our data preprocessing pipeline. At 500 Hz frame rates, we demonstrate a 75% increase in PNR using the denoised pHiLo reconstruction compared to pWF. Utilizing DeepCAD-RT, we show significant PNR improvements using both structured illumination (SI) reconstruction methods with OS-SIM showing a 59% increase in PNR after denoising. Both pHiLo and OS-SIM reconstructions result in a ≈65% decrease in the mean correlation coefficient of the ΔF/F time courses between ROIs in comparison with pWF, indicating the potential to remove background fluorescent transients from out-of-focus cells.

Advancing Multi‐Optical Performances in Lanthanide‐Doped Fluoride Core@Shell Nanoarchitectures by Minimizing Cation Intermixing

AbstractCore/shell structures are widely employed to enhance photoluminescence efficiency and manipulate excitation dynamics in lanthanide‐doped fluoride nanoparticles (NPs). However, prominent cation intermixing leads to significant deviations from the expected optical performances, presenting a formidable challenge in a deeper understanding of the chemical processes involved, as well as developing efficient suppression methods. Here, a reliable and facile multi‐optical reference strategy to reveal the integral cation intermixing processes is designed. This strategy analyzes the presence of Ce3+ ions in the core (or shell) and their influences on near‐infrared light‐triggered upconversion (UC), ultraviolet (UV)‐activated downshifting (DS), and X‐ray‐excited optical/persistent luminescence (XEOL/XEPL) of Tb3+ ions in the shell (or core). The results demonstrate that a thin surface layer of core NPs dissolves to reach dissolution equilibrium, which then distributes throughout the entire shell layer, rather than being confined to an interfacial region. It is further developed an improved technique to greatly inhibit cation intermixing by successive shell growth with excessive cation precursors, enabling superior multi‐optical performances, including UC, DS, XEOL/XEPL, and time‐dependent multicolor evolution. The findings significantly advance the development of lanthanide‐doped fluoride core/shell NPs with superior optical performances, broadening their potential applications in bio‐medicine, healthcare, industrial inspection, and optical information science.

Noise & mottle suppression methods for cumulative Cherenkov images of radiation therapy delivery.

Cherenkov imaging during radiotherapy provides a real time visualization of beam delivery on patient tissue, which can be used dynamically for incident detection or to review a summary of the delivered surface signal for treatment verification. Very few photons form the images, and one limitation is that the noise level per frame can be quite high, and mottle in the cumulative processed images can cause mild overall noise. This work focused on removing or suppressing noise via image postprocessing. Images were analyzed for peak-signal-to-noise and spatial frequencies present, and several established noise/mottle reduction algorithms were chosen based upon these observations. These included total variation minimization (TV-L1), non-local means filter (NLM), block-matching 3D (BM3D), alpha (adaptive) trimmed mean (ATM), and bilateral filtering. Each were applied to images acquired using a BeamSite camera (DoseOptics) imaged signal from 6x photons from a TrueBeam linac delivering dose at 600 MU/min incident on an anthropomorphic phantom and tissue slab phantom in various configurations and beam angles. The standard denoised images were tested for PSNR, noise power spectrum (NPS) and image sharpness. The average peak-signal-to-noise ratio (PSNR) increase was 17.4% for TV-L1. NLM denoising increased the average PSNR by 19.1%, BM3D processing increased it by12.1% and the bilateral filter increased the average PSNR by 19.0%. Lastly, the ATM filter resulted in the lowest average PSNR increase of 10.9%. Of all of these, the NLM and bilateral filters produced improved edge sharpness with, generally, the lowest NPS curve. For cumulative image Cherenkov data, NLM and the bilateral filter yielded optimal denoising with the TV-L1 algorithm giving comparable results. Single video frame Cherenkov images exhibit much higher noise levels compared to cumulative images. Noise suppression algorithms for these frame rates will likely be a different processing pipeline involving these filters incorporated with machine learning.

Sinomenine hydrochloride alleviates autoimmune myocarditis via suppressing Th1 cell induced-M1 macrophage pyroptosis

Abstract Background Myocarditis is typical of the complex inflammatory response in the heart with high morbidity and mortality, accompanied by tissue damage and cardiac fibrosis. Previous studies have demonstrated that CD4+ T cells are crucial to the cardiac autoimmunity of myocarditis, but the role of different CD4+ T cell subsets remains unclear. Recently, sinomenine hydrochloride, a natural compound from the traditional Chinese herb Sinomenium acutum, is reported to inhibit T cell proliferation in some systemic autoimmune diseases such as rheumatoid arthritis. Given that autoimmune myocarditis is an organ-specific autoimmune disease, the effect of sinomenine hydrochloride on autoimmune myocarditis demands further study. Purpose This study aims to elucidate the role of CD4+ T-helper 1 (Th1) cells in regulating inflammatory cell death in autoimmune myocarditis and investigate the pharmacological effect of sinomenine hydrochloride on CD4+ Th1 cells and autoimmune myocarditis. Methods Male Balb/c mice were immunized with myosin heavy chain-α peptides to establish the experimental autoimmune myocarditis (EAM) model, followed by the treatments with sinomenine hydrochloride by gavage. Tohoku Hospital Pediatrics-1 (THP-1) cell line and murine T cells were used for in vitro experiments. Results In the acute phase of EAM, we observed plenty of CD4+ T cells infiltrating the heart tissue. RNA sequencing revealed that among the genes encoding typical transcription factors of CD4+ T cell subset, only Tbx21 and Stat4, both representing Th1 cells, were upregulated, which consisted with immunofluorescence staining. Further studies indicated that Th1 cells not only mediated classical activated macrophage (M1 macrophage) polarization by secreting interferon-γ (IFN-γ) but also associated with pro-inflammatory macrophage pyroptosis in the EAM, with the cleavage of gasdermin D (GSDMD), caspase-1 and IL-1β. Subsequent in vitro experiments confirmed that IFN-γ could dramatically promote macrophage pyroptosis combined with ATP, leading to the release of IL-1β and IL-18(Figure 1). To find an effective method for Th1 cell suppression and myocarditis remission, we identified that sinomenine hydrochloride impeded Th1 cell activation, differentiation and proliferation in vitro. Furthermore, sinomenine hydrochloride inhibited Th1 cell infiltration and M1 macrophage pyroptosis in the EAM heart, In the chronic phase, cardiac remodeling and fibrosis were suppressed, and cardiac systolic function was significantly improved in the treatment group (Figure 2). Conclusion Our study demonstrates that Th1 cells-induced M1 macrophage pyroptosis is an essential pathogenic mechanism of EAM. Sinomenine hydrochloride can alleviate the severity and progression of EAM, which may become a novel therapeutic strategy for treating myocarditis.Figure 1Figure 2

Nonstationary adaptive S-wave leakage suppression of ocean-bottom node data

Ocean-bottom nodes (OBNs) are widely used because of their wide azimuth, long-offset, and low-frequency advantages. However, in the vertical component of the OBN geophone, a significant amount of S-wave induced noise may be recorded. This significantly impacts the quality of the vertical component data and may affect the follow-up merging of dual-sensor data. Some commercially available methods apply normal-moveout correction, which requires velocity data. We avoid this with an adaptive matching method, which relies solely on seismic data. Therefore, we develop a novel adaptive subtraction method for S-wave leakage suppression using the horizontal component as the noise model. Our method effectively handles the nonstationarity of the input seismic data in all time and space directions, mitigating instability caused by manual selection of the smoothing radius. A method is introduced for estimating a global nonstationary smoothing radius using the noise model. Compared with commercial and stationary smoothing methods, our method can better suppress S-wave noise while balancing residual noise and signal leakage more effectively. Synthetic and field data examples demonstrate significant improvement with our method.

Research on anti-swing control system of slewing crane based on fuzzy PID.

The slewing crane is easily affected by the wind and the manipulation level of the operator when it is working, which in turn impacts its swing angle, affects the working efficiency and safety of the crane. Toimprove the operation safety andreduce the swing angle, this study design and research the anti-swing control system for the slewing crane. Firstly, based on the Lagrange equation, the dynamic model of the crane hoisting system is established. Then, the mechanical anti-swing mechanism driven by a four-motor rope is established. Finally, based on fuzzy PID control, an anti-swing platform is established, and the performance of anti-swing control is verified. Compared to no anti-swing control, the in-plane swing angle θ1 and out-of-plane swing angle θ2 of the lifting arm are reduced by 88% and 75% respectively. The results show that the four-motor rope-driven anti-swing mechanism can achieve better anti-swing effect, which provides an effective swing suppression method for the slewing crane in the actual operation.

Millimeter-Wave Radar Clutter Suppression Based on Cycle-Consistency Generative Adversarial Network

Vehicle-mounted millimeter-wave radar is widely used in autonomous driving systems for its ability to observe road scenes at all times and in all weathers. However, the data collected by millimeter-wave radar are seriously affected by the existence of clutter. This clutter will result in false detection during object detection. To address this issue, a feature extraction network with clutter suppression is necessary. This paper proposes a new clutter suppression method for millimeter-wave Range–Angle (RA) images based on a cycle-consistency generative adversarial network (CycleGAN). The generator of the method can be used as the feature extraction network of the object detection. The method aims to convert cluttered images into clutter-free images by unsupervised learning. In this method, an attention gate (AG) is introduced into the generator, a spatial attention mechanism that improves the ability of the model to automatically learn to focus on the features of targets and suppress the clutter of the background. Additionally, the target consistency loss term is added to the loss function to maintain target integrity while suppressing network training overfitting. The public dataset CRUW is utilized to evaluate the performance of the proposed method, which is compared and analyzed with traditional methods and deep learning methods. Experimental results show that the peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) of the proposed method reach 39.846 and 0.990, respectively.

Suppression of quantum dissipation: a cooperative effect of quantum squeezing and quantum measurement.

The ability to isolate a quantum system from its environment is of fundamental interest and importance in optical quantum science and technology. Here we propose an experimentally feasible scheme for beating environment-induced dissipation in an open two-level system coupled to a parametrically driven cavity. The mechanism relies on a novel, to the best of our knowledge, cooperation between light-matter coupling enhancement and frequent measurements. We demonstrate that, in the presence of the cooperation, the system dynamics can be completely dominated by the effective system-cavity interaction, and the dissipative effects from the system-environment coupling can be surprisingly ignored. This work provides a generic method of dissipation suppression in a variety of quantum mechanical platforms, including natural atoms and superconducting circuits.

Oscillation Suppression of Grid-Following Converters by Grid-Forming Converters with Adaptive Droop Control

The high penetration of renewable energy sources (RESs) and power electronics devices has led to a continuous decline in power system stability. Due to the instability of grid-following converters (GFLCs) in weak grids, the grid-forming converters (GFMCs) have gained widespread attention featuring their flexible frequency and voltage regulation capabilities, as well as the satisfactory grid-supporting services, such as inertia and damping, et al. Notably, the risk of wideband oscillations in modern power grids is increasingly exacerbated by the reduced number of synchronous generators (SGs). Thus, the wideband oscillation suppression method based on adaptive active power droop control of GFMCs is presented in this paper. First, the stability of the hybrid grid-forming and grid-following system is obtained according to the improved short circuit ratio (ISCR), where the GFMC is in parallel at the point of common coupling (PCC) of the GFLC. Then, an adaptive adjustment strategy of the active power droop control is proposed to enhance the oscillation suppression capability across the full frequency range, thereby mitigating the wideband oscillation caused by phase-locked loop (PLL) synchronization in the GFLCs. Additionally, a first-order inertia control unit is added to the active and reactive power droop controllers to mitigate frequency and voltage variations as well as suppress potential mid-to-high frequency resonance. Finally, the wideband oscillation suppression strategy is validated by the simulation and experimental results.

Improvement of Criminisi’s Stripe Noise Suppression Method for Side-Scan Sonar Images

In response to the problem of stripe noise significantly reducing the clarity and details of side-scan sonar images due to various factors, the authors of this paper propose an improved Criminisi method for stripe noise suppression. To address the issues encountered in the Criminisi algorithm during the suppression of stripe noise in side-scan sonar images, the following steps are suggested: firstly, introduce dynamic weights in the priority calculation to adaptively adjust the confidence and data term weights based on the current patch’s texture complexity; secondly, utilize the Sobel operator in the data term calculation to capture the image edge information more accurately; and, thirdly, optimize the matching block search process by introducing the Manhattan distance in addition to the Sum of Squared Differences (SSD) criterion to further select the best matching block while transitioning from a global search to a local search. Experimental validation was conducted using simulated stripe noise images, comparing the proposed method with four traditional denoising techniques. The results demonstrate that the denoising effectiveness of the proposed method is superior, effectively restoring texture in noisy regions while preserving texture structure integrity. Ablation experiments validate the effectiveness of the proposed improvements. Denoising experiments on real noisy images show satisfactory results with this method, and combining it with Fourier transform for additional smoothing in certain cases may yield even better results.

Force mechanism analysis of composite microbial dust suppressants based on extracellular polymeric substances (EPS) mode components

As an important potential dust suppression method, the slow onset time is one of the key factors that restrict the effect of microbial dust suppressant. In the early stage, we have confirmed that extracellular polymeric substances (EPS) can improve the dust suppression effect by wetting coal dust and increasing Ca2+ nucleation sites. Therefore, in this study, chitosan (CTS) and bovine serum albumin (BSA) in different ratios (CTS: BSA = 1:1, 1:2, 2:1) as model molecules of EPS were combined with Bacillus subtilis to prepare efficient and fast microbial dust suppressants. Furthermore, the interaction forces were analyzed through molecular dynamics simulation. Results showed that adding CTS and BSA would improve the dust suppression effect, and the dust suppression effect was the best when the ratio of CTS: BSA was 1:2. In addition, the contact angle decreased as the BSA content increased. The Fourier transform infrared spectroscopy (FTIR) results showed that when the ratio of CTS to BSA was 1:2, the dust suppressants were easier to interact with coal dust by the key functional groups and form calcite type CaCO3. The molecular dynamics simulation results showed that the main interaction was Van der Waals force. In addition, the interaction force was strongest when CTS: BSA was 1:2, increasing by 137% compared with the microbial dust suppressants without CTS or BSA. This study provides theoretical support for the development of efficient and rapid microbial dust suppressants.

Research on subsurface damage mechanism and suppression method of ultrasonic vibration–assisted grinding of sapphire components under extreme service environment

Research on subsurface damage mechanism and suppression method of ultrasonic vibration–assisted grinding of sapphire components under extreme service environment

A spurious-free SAW filter employing a novel transducer structure on bulk LiNbO3 substrates

In order to cope with the ubiquitous wireless connections and the rapid transmission of data in communication systems, surface acoustic wave filters are required to have high performance. With a large electromechanical coupling coefficient (K2), conventional bulk 64°YX-LiNbO3 substrates are attractive for low-cost mass production of surface acoustic wave (SAW) resonators. However, the influence of the transverse modes restricts their practical application. In this work, a novel and simple tilted transducer is fabricated, and successfully achieved spurious-free passband SAW filters based on 64°YX-LN piezoelectric substrates. In this work, the transverse modes suppression method is theoretically studied and designed, and the effectiveness of the energy concentrated structure is verified. The optimal tilt IDT angle for eliminating the influence of transverse modes in Al/64 °YX-LN structure is about 6°. Subsequently, based on the above research, three types of tilted resonators were designed and fabricated, which not only verified the reliability of the aforementioned results but also enhanced the performance of the devices, compensating for the deficiency caused by the reduction of the Q factor due to the tilted structure. Finally, based on the double busbar tilted structure, a SAW filter was designed, which has a large 3 dB fractional bandwidth (FBW) and a spurious-free passband. The SAW filter designed in this work has low cost and excellent performance, and can be mass produced. In the field of low-cost RF devices, this work contributes to maintaining the competitive advantage of SAW technology.

Significant feature suppression and cross-feature fusion networks for fine-grained visual classification

The technique of extracting different distinguishing features by locating different part regions to achieve fine-grained visual classification (FGVC) has made significant improvements. Utilizing attention mechanisms for feature extraction has become one of the mainstream methods in computer vision, but these methods have certain limitations. They typically focus on the most discriminative regions and directly combine the features of these parts, neglecting other less prominent yet still discriminative regions. Additionally, these methods may not fully explore the intrinsic connections between higher-order and lower-order features to optimize model classification performance. By considering the potential relationships between different higher-order feature representations in the object image, we can enable the integrated higher-order features to contribute more significantly to the model’s classification decision-making capabilities. To this end, we propose a saliency feature suppression and cross-feature fusion network model (SFSCF-Net) to explore the interaction learning between different higher-order feature representations. These include (1) an object-level image generator (OIG): the intersection of the output feature maps of the last two convolutional blocks of the backbone network is used as an object mask and mapped to the original image for cropping to obtain an object-level image, which can effectively reduce the interference caused by complex backgrounds. (2) A saliency feature suppression module (SFSM): the most distinguishing part of the object image is obtained by a feature extractor, and the part is masked by a two-dimensional suppression method, which improves the accuracy of feature suppression. (3) A cross-feature fusion method (CFM) based on inter-layer interaction: the output feature maps of different network layers are interactively integrated to obtain high-dimensional features, and then the high-dimensional features are channel compressed to obtain the inter-layer interaction feature representation, which enriches the output feature semantic information. The proposed SFSCF-Net can be trained end-to-end and achieves state-of-the-art or competitive results on four FGVC benchmark datasets.