Structural Noise Research Articles

DIABETIC RETINOPATHY SEVERITY CLASSIFICATION USING CUSTOMIZED CNN

Diabetic retinopathy (DR) is caused by diabetes, and could lead to permanent blindness. Diabetes causes damage to the arteries and veins of eye and further results in the loss of vision. Detection of DR is challenging as there are not many symptoms of the disease at the early stage. Opthalmologists can detect the DR, and analyze the severity by visual analysis of the fundus images. The population of diabetes-affected people is large and hence the manual detection and analysis is tedious and expensive. The automated screening and severity detection systems are therefore needed. In this paper, the experiments with CNN for classifying the DR images into five classes (NO DR, mild, moderate, severe and PDR) are conducted on the 2015 and 2019 Blindness Detection Kaggle dataset. Customized CNN model is designed in order to provide accurate severity classification. The CNN model consists of two convolutional layers, two max-pooling layers, one flattening layer and two dense layers. The retinal fundus images present structural and impulsive noise. Gaussian blur technique is applied as preprocessing to reduce the noise. We experimented with Adam’s optimizer as well as SGD optimizer and observed that the best classification results obtained using Adam’s optimizer (89% accuracy) are promising but warrant further investigation. Moreover, our approach has been able to have good and consistent AUC for all the classes, and particularly worked better than the existing approaches for the detection of class 3 (severe DR).

Low-frequency vibration reduction of an underwater metamaterial plate excited by a turbulent boundary layer

Flow-induced structural noise is an important component of hydrodynamic noise of underwater structures. Local resonance metamaterials are considered to have excellent performance and enormous potential in the field of low-frequency vibration and noise control. To verify its potential, the paper derived the underwater band gap of a lateral local resonance (LLR) plate through the plane wave expansion (PWE). Then, utilizing the modal superposition approach and Rayleigh integral technique, the vibro-acoustic response of a LLR plate under a turbulent boundary layer (TBL) excitation is obtained. Finite element certification is also conducted through an uncorrelated wall plane wave technique. Parametric study is conducted to analyse the factors which influence the control effects. The result shows that the plate exhibits excellent suppression performance for flow-induced vibration at band gap frequencies. The band gaps and suppression ranges generated by the underwater metamaterial plate, are dramatically narrowed due to the thick fluid load. The paper provides theoretical guidance for the control of flow-induced structural vibration and the application of acoustic metamaterials.

Analyzing Robustness of Angluin's L$^*$ Algorithm in Presence of Noise

Angluin's L$^*$ algorithm learns the minimal deterministic finite automaton (DFA) of a regular language using membership and equivalence queries. Its probabilistic approximatively correct (PAC) version substitutes an equivalence query by numerous random membership queries to get a high level confidence to the answer. Thus it can be applied to any kind of device and may be viewed as an algorithm for synthesizing an automaton abstracting the behavior of the device based on observations. Here we are interested on how Angluin's PAC learning algorithm behaves for devices which are obtained from a DFA by introducing some noise. More precisely we study whether Angluin's algorithm reduces the noise and produces a DFA closer to the original one than the noisy device. We propose several ways to introduce the noise: (1) the noisy device inverts the classification of words w.r.t. the DFA with a small probability, (2) the noisy device modifies with a small probability the letters of the word before asking its classification w.r.t. the DFA, (3) the noisy device combines the classification of a word w.r.t. the DFA and its classification w.r.t. a counter automaton, and (4) the noisy DFA is obtained by a random process from two DFA such that the language of the first one is included in the second one. Then when a word is accepted (resp. rejected) by the first (resp. second) one, it is also accepted (resp. rejected) and in the remaining cases, it is accepted with probability 0.5. Our main experimental contributions consist in showing that: (1) Angluin's algorithm behaves well whenever the noisy device is produced by a random process, (2) but poorly with a structured noise, and, that (3) is able to eliminate pathological behaviours specified in a regular way. Theoretically, we show that randomness almost surely yields systems with non-recursively enumerable languages.

Joint contrastive learning of structural and semantic for graph collaborative filtering

Recently, graph collaborative filtering has been proposed as a superior recommendation technique, due to its outstanding capability of capturing high-order correlations from user–item interactions. Despite effectiveness, it still suffers from two limitations on representation learning: (1) imbalanced data distribution. User/item nodes with high-degree usually interfere with the representation learning process of low-degree nodes, deteriorating the recommendation accuracy; (2) structural noise vulnerability. The neighborhood aggregation scheme for the node representing could easily amplify structural noises, damaging the robustness of recommenders. The latest studies have made progress in overcoming the two limitations by capturing self-supervision signals via contrastive learning. Nevertheless, they somehow rely on heuristics to design, at a certain distance from a more comprehensive settlement.To make contrastive learning more entrenched with graph collaborative filtering, this paper proposes a multi-task framework JCG that integrates joint contrastive learning strategies with a backbone of graph-based recommender. Specifically, (1) we design six contrastive learning tasks categorized as structural and semantic, and obtain nine joint strategies by combining two tasks from each category; (2) we systematically study the effect of joint strategies on graph collaborative filtering, finding that JCG with feasible strategies could achieve up to 18.5% and 17.9% performance gain over the backbone on datasets of Yelp and ML, respectively; (3) performance comparison with other competitive baselines demonstrates the superiority of our JCG, especially that on mitigating the two limitations of data imbalance and noise vulnerability. We hope that our conclusion could provide new insight for future studies.

Spectral collocation method for free vibration of sandwich plates containing a viscoelastic core

Viscoelastic constrained layer damping is a simple and efficient method for reducing noise, vibration, and fatigue in metal structures. In this work, a spectral collocation method utilizing a layer-wise plate theory was established to inspect the vibration characteristics of a sandwich plate with a viscoelastic core. The displacements of each layer satisfied the Mindlin plate theory. The core's transverse shear stress remained constant. The three layers’ transverse displacements were assumed to be identical. The displacement fields were reduced to nine variables by utilizing the interfaces’ continuity of the displacements. The equations of motion were obtained by utilizing Hamilton’s principle. The viscoelastic core’s material properties were considered as frequency-dependent. To address the complex eigenvalue problem, an iterative algorithm was used. The theoretical results were compared with published theoretical and experimental data for fully clamped rectangular sandwich plates to validate the proposed method. The modified Oberst beam method was applied by fitting the impulse response function to obtain the adhesive’s frequency-dependent storage modulus and loss factor. The natural frequencies and corresponding loss factors of three sandwich plates with different adhesives were measured by conducting impact tests. The numerical results from the present theoretical method agreed well with the measured results.

Combining iterative modal strain energy method with BEM-SEA to predict structural noise from railway composite bridge damped with CLD

ABSTRACT This paper presents a combined method to predict structural noise from railway composite bridge damped with constrained layer damping (CLD), in which the frequency-dependent properties of the viscoelastic material are accurately considered through the iterative modal strain energy (IMSE) method. To efficiently determine the forces transmitted to bridge, a simplified vehicle-track coupling model is proposed. Taking 200 Hz as the boundary frequency, the IMSE-boundary element method (BEM) and the IMSE-statistical energy analysis (SEA) are used in the frequency range of 20 ~ 200 Hz and 200 ~ 2000 Hz to realize the full-band noise prediction. Then, the on-site noise tests of a three-span railway composite bridge are carried out to verify the proposed method. Finally, the influences of the key design parameters of CLD on noise reduction are discussed, so as to provide reference for the research and application of CLD in railway bridge engineering.

The Graduate Program in Acoustics at Penn State

The Graduate Program in Acoustics at Penn State offers graduate degrees (M.Eng., M.S., Ph.D.) in Acoustics, with courses and research opportunities in a wide variety of subfields. Our 820 alumni are employed around the world in military and government labs, academic institutions, consulting firms, and consumer audio and related industries. Our 40 + faculty from several disciplines conduct research and teach courses in structural acoustics, nonlinear acoustics, architectural acoustics, signal processing, aeroacoustics, biomedical ultrasound, transducers, computational acoustics, noise and vibration control, acoustic metamaterials, psychoacoustics, and underwater acoustics. Course offerings include fundamentals of acoustics and vibration, electroacoustic transducers, signal processing, acoustics in fluid media, sound and structure interaction, digital signal processing, experimental techniques, acoustic measurements and data analysis, ocean acoustics, architectural acoustics, noise control engineering, nonlinear acoustics, outdoor sound propagation, computational acoustics, biomedical ultrasound, flow induced noise, spatial sound and three-dimensional audio, and the acoustics of musical instruments. Distance education students pursuing the M.Eng. degree join resident students in a hybrid classroom environment. This poster highlights faculty research areas, laboratory facilities, student demographics, successful graduates, and recent enrollment and employment trends.

Ultrasonic metrics for large-area rapid wrinkle detection and classification in composites

Due to their high strength-to-weight ratio, composite materials are now in use in many high-stress applications, particularly where light weight is also a requirement. In these situations, the detrimental knock-down in mechanical strength due to an out-of-plane wrinkle defect can have serious consequences and is the reason for a requirement to rapidly detect any such wrinkles at manufacture. Unfortunately, current ultrasonic inspection techniques used for quality control at manufacture are not sensitive enough to detect these wrinkles above coherent structural noise variations. This paper exploits the ply resonance that is a characteristic of multi-layer structures to generate two new metrics for both detection and classification of out-of-plane wrinkles, due to their perturbations of the ply spacing. These can be measured at every location on a structure using the instantaneous frequency, which is the rate of change of phase in the pulse-echo ultrasonic response. The proposed two new metrics for detection and classification of wrinkles are mean spacing and spacing difference and they can be applied to each waveform in real time, as it is acquired. Use of an analytical model to predict the ultrasonic response of the structure has allowed an understanding of how these metrics will be affected by various wrinkle types and how they can not only detect wrinkles but also classify the type of wrinkle and provide an approximate indication of severity. Three main types of wrinkle are considered: classic wrinkles near the mid-plane of a structure, back-surface wrinkles formed from a resin bulge near the back of a structure and folded wrinkles where several plies can be folded over completely in the bulk of the structure. Both simulations and experimental results demonstrate the effectiveness of these metrics on various types of structure, including carbon-fibre and hybrid carbon/glass-fibre composites with a range of ply thicknesses and wrinkle types.

Joint first and second order total variation decomposition for remote sensing images destriping

ABSTRACT Stripe noise remains a significant source of errors and image quality degradation in remote sensing systems. A prominent approach for tackling this problem is the first-order Total Variation (TV) regularization, which has a proven efficiency in dealing with stripe noise. Unfortunately, denoised images, in this case, may suffer from texture loss and staircase artefacts around smooth areas. In this paper, a novel image decomposition scheme is proposed to tackle these drawbacks. This decomposition is based on the use of directional first- and second-order TV regularizers that are employed to separate the clear image from the stripe component while considering the directionality and smoothness of the latter. The proposed model is solved using a Chambolle-based algorithm and its performance is compared to traditional destriping methods using different noise structures and intensities. The results have shown comparable performance to the existing state-of-the-art methods with some improvements in structure preservation and noise cancellation.

Numerical Investigation of the Flow Structure and Acoustical Noise in a Suppressor

The noise created when a firearm is fired has a lot of adverse effects on humans and the environment, so analyzing and attenuating this noise is essential. This study aimed to examine propellant flow and the sound generated by this flow using a hybrid computational fluid dynamics and computational aeroacoustics method. The compatibility of this numerical study was also validated by comparing it with the experimental result. The impact of critical parameters, such as the shape and number of baffles of the suppressor, was studied. Finally, the overpressure and acoustics results of different suppressors were compared with each other and with the unsuppressed condition. According to the result, the suppressor with curved baffles shows a better performance. When the exit pressure at the tip of suppressor compared to the initial inlet pressure to the suppressor the percentage drop in overpressure was 76.01%, 78.79% and 81.3% for the suppressor with one, three and five curved baffles, respectively. For condition without suppressor, 169.49[Formula: see text]dB of SPL was recorded. When using a suppressor without a baffle, this value was reduced to 162.13[Formula: see text]dB. For the suppressor with one, three and five curved baffles, the SPL value was 160.234[Formula: see text]dB, 159.43[Formula: see text]dB and 158.11[Formula: see text]dB, respectively. Generally, this study shows that the suppressor’s efficiency is directly proportional to its shape and number of baffles.

Learning CT-free attenuation-corrected total-body PET images through deep learning.

Total-body PET/CT scanners with long axial fields of view have enabled unprecedented image quality and quantitative accuracy. However, the ionizing radiation from CT is a major issue in PET imaging, which is more evident with reduced radiopharmaceutical doses in total-body PET/CT. Therefore, we attempted to generate CT-free attenuation-corrected (CTF-AC) total-body PET images through deep learning. Based on total-body PET data from 122 subjects (29 females and 93 males), a well-established cycle-consistent generative adversarial network (Cycle-GAN) was employed to generate CTF-AC total-body PET images directly while introducing site structures as prior information. Statistical analyses, including Pearson correlation coefficient (PCC) and t-tests, were utilized for the correlation measurements. The generated CTF-AC total-body PET images closely resembled real AC PET images, showing reduced noise and good contrast in different tissue structures. The obtained peak signal-to-noise ratio and structural similarity index measure values were 36.92 ± 5.49 dB (p < 0.01) and 0.980 ± 0.041 (p < 0.01), respectively. Furthermore, the standardized uptake value (SUV) distribution was consistent with that of real AC PET images. Our approach could directly generate CTF-AC total-body PET images, greatly reducing the radiation risk to patients from redundant anatomical examinations. Moreover, the model was validated based on a multidose-level NAC-AC PET dataset, demonstrating the potential of our method for low-dose PET attenuation correction. In future work, we will attempt to validate the proposed method with total-body PET/CT systems in more clinical practices. The ionizing radiation from CT is a major issue in PET imaging, which is more evident with reduced radiopharmaceutical doses in total-body PET/CT. Our CT-free PET attenuation correction method would be beneficial for a wide range of patient populations, especially for pediatric examinations and patients who need multiple scans or who require long-term follow-up. • CT is the main source of radiation in PET/CT imaging, especially for total-body PET/CT devices, and reduced radiopharmaceutical doses make the radiation burden from CT more obvious. • The CT-free PET attenuation correction method would be beneficial for patients who need multiple scans or long-term follow-up by reducing additional radiation from redundant anatomical examinations. • The proposed method could directly generate CT-free attenuation-corrected (CTF-AC) total-body PET images, which is beneficial for PET/MRI or PET-only devices lacking CT image poses.

Inspection of antimicrobial particles in milk using RGB-laser scattering imaging combined with chemometric procedures

The immobilisation of phytochemicals on inert particles has demonstrated high potential to develop new formulations of natural antimicrobials for preserving liquid foods such as milk. This work aimed to test the capability of the laser scattering imaging technique combined with chemometric procedures to inspect antimicrobial particles in milk matrices non-destructively. The RGB laser's interactions with the milk matrix were captured as diffraction patterns in digital images, from which data were extracted after imaging processing and modelled by chemometric procedures. The antimicrobials were based on solid inorganic particles (silica) functionalised with a plant essential oil (eugenol). The effect of the natural variability from five types of milk was studied (cow 0, 1.5 and 3.5% lipids, goat and sheep). The inspection functions focused on predicting the type of milk, presence of antimicrobial particles, functionalisation of particles and type of inorganic structure. PLS-DA was applied to perform classification models for each property prediction, combined with pre-processing and variable selection procedures to isolate the variability of the property of interest. The results revealed a high classification capacity for milk types. The properties of antimicrobials were also successfully predicted. In this case, data pre-processing and variable selection procedures had to be applied to reduce structural noise and improve the models from raw data. Combining imaging and chemometrics techniques could capture and model the variability generated by antimicrobial particles independently of the nature of milk. It established the basis for developing a possible non-destructive inspection system to control liquid food containing these new antimicrobials.

Temperature effects on noise radiated by concrete railway structures

In recent years, concerns have arisen regarding the periodic deformation of high-speed railway tracks owing to the temperature effects caused by the solar radiation. This phenomenon has prompted investigations into temperature-induced changes in the dynamic responses of vehicle–track–bridge systems. Previous studies have commonly taken the additional track irregularity caused by thermal deformation as the sole factor affecting dynamic responses while neglected changes in the structural frequency response characteristics. Structural radiation noise is also a prominent issue that affects the lives of people and the environment. However, studies on track–bridge-borne noise rarely consider the influence of temperature. To address the impact of temperature on structure-borne noise radiated by tracks and bridges, this study employed a sequentially coupled multi-physics analysis method. This approach allowed a comprehensive exploration of the temperature effects on track irregularities, structural frequency responses, and sound radiation efficiency. A full-scale temperature deformation field test of a CRTS III slab was also conducted to determine the necessary initial thermodynamic boundary conditions. The analysis results revealed that the thermally induced deterioration of track irregularity and changes in structural frequency response characteristics were the primary factors affecting structure-borne noise radiated by track–bridges under local climate conditions.

The flow-induced structural vibration noise suppression mechanism of a wing–plate model by the junction suction and trailing edge blowing

At medium to high speeds, the flow-induced structural vibration noise generated by structural vibration under the excitation of the turbulent fluctuating pressure can seriously damage the acoustic performance of underwater vehicles. In this paper, we have proposed a new and noiseless method, called junction suction and trailing edge blowing (JSTB), to reduce the flow-induced structural vibration noise of the wing–plate model. We have investigated the effects of JSTB on the flow and sound fields through a combination of numerical simulations and experimental tests. The experiments are conducted in a low-noise gravity water tunnel. Numerical results show that JSTB significantly reduces the peak values of the internal self-noise at 70 Hz, 180 Hz, 225 Hz, and 265 Hz when the velocity is 6.87ms−1. The model radiated noise reduction is more prominent in the 20–1000 Hz frequency range, with an optimal noise reduction level of 15.4 dB. The simulation results agree well with the experimental results. JSTB can effectively reduce the internal sound pressure level, turbulent fluctuating pressure, and radiated sound power of the model. This work provides a novel solution for the flow-induced structural vibration noise control of underwater vehicles.

Dynamic Receptance Analysis Combined with Hybrid FE-SEA Method to Predict Structural Noise from Long-Span Cable-Stayed Bridge in Urban Rail Transit

In this study, dynamic receptance analysis (DRA) is proposed and combined with hybrid finite element (FE)-statistical energy analysis (SEA) method to accurately predict structural noise from long-span cable-stayed bridge (LSCB) with steel box composite girder (SBCG) in urban rail transit (URT). To begin with, a vertical vehicle–track coupling model in frequency domain is established based on DRA, in which the rail is represented by an infinite Timoshenko beam supported by a series of fasteners that are regarded as springs with complex stiffness. The floating slab is regarded as the Euler beam with both ends free supported by steel springs. Using this model, the spectrums of the wheel–rail force and the forces transferred to the bridge can be efficiently obtained by taking rail roughness as the excitation. Due to the low modal density of the concrete deck, the hybrid FE-SEA method is introduced to establish the noise prediction model, in which the discontinuity caused by using distinct models for different frequency bands is avoided. Then the on-site noise tests of a LSCB with SBCG in URT are carried out to verify of the proposed method. Finally, based on the prediction method, the acoustic contributions of the bridge components are analyzed in detail. The force transfer characteristics as well as the noise reduction effects of different track structures are thoroughly investigated, so as to provide reference for the future research on bridge-borne noise control.

Design and Testing of a New Type of Planetary Traction Drive Bearing-Type Reducer

This paper presents the design and development of a new type of planetary traction drive bearing-type reducer. In this design, the transmission outer ring is replaced with an elastic ring. The design constructs a circular arc at the axial end of the rolling body’s contour line. This ensures that the contact point of this arc with the reducer’s outer ring and the inner ring’s axial end face is maintained on the radial traction contact line. As a result, it can replace the thrust bearing and provide an axial support function. It has the advantages of simple structure, easy processing, smooth transmission, and low noise. This paper first introduces the design and development process of this bearing-type reducer and presents systematic research on its transmission principle and dynamics. Subsequently, in response to the edge effect phenomenon of the outer ring contact line, the contour line of the outer ring is refined by adopting the shaping method used for bearing rollers, establishing a full circular arc profile shaping method, which significantly improves its edge effect. Finally, in our investigations, combined with experimental tests, a prototype of the bearing-type reducer was fabricated, and the speed ratio, torque, and transmission efficiency of the reducer were studied. The results demonstrate that the bearing-type reducer can achieve high transmission accuracy and efficiency. The transmission performance varies significantly under different lubrication conditions, with the peak efficiency reaching as high as 99.97% when using Santotrac 50 traction oil. The results verify the feasibility of the proposed design method and have the potential to be applied in wheel hub motors and robot joints.

Numerical study on aerodynamic performance and noise of protrusions structures near the trailing edge

Based on the NACA0012 airfoil, different types of protrusions namely serrated protrusions and arc-shaped protrusions are set near the trailing edge. Both new airfoils have a certain decrease in flow performance however there is a significant reduction in aerodynamic noise of airfoils. This is because the protrusions damage the vortex near the trailing edge. In addition, the noise reduction of serrated protrusions is more significant than t of arc-shaped protrusions, about 2.7dB.

Does Noise Really Matter? Investigation into the Influence of Noisy Labels on BERT-Based Question Answering System

Recent works with the BERT-based models demonstrate their generalization ability and high performance on the new domain tasks. However, this kind of model requires a large amount of data. Collecting this data can be error-prone, and it is important to know: how the errors in data affect the quality of the model. In this work, we study the impact of data with different errors – noisy data on the training of the question answering-over-text BERT-model. We use the concept of random, structural and irrelevant question noises. We study the robustness of QAT models during the training process with different settings, datasets and noise types and discuss possible reasons. We also propose a real-world domain dataset to probe our findings in a real-world scenario. The results of an experimental study showed that following developed recommendations allowed performance improvement up to 3.6% in a real-world setting.

Computational and Theoretical Investigation of Acoustical and Vibrational Properties of Rigid Thin Material

A computational and theoretical investigation of acoustical and vibrational properties of rigid thin fiberglass material was carried out for different boundary conditions. Fiberglass materials could be applied in industries varying from the aircraft and automotive sectors to the built environment and construction sectors. Plate vibration and acoustic radiation were applied to predict the deflection of the thin fiberglass material and sound radiation efficiency at different locations on its surface, while a study-controlled equation of motion known as the Kirchhoff thin plate theory was applied for a COMSOL simulation of the thin material to determine the deflection of the plate and to obtain stress distribution, velocity contour, displacement, and acoustic pressure at the first resonance of the material. The results of this paper show that thin fiberglass material could be applied to sandwich building elements to form panels for reducing airborne noise and to lessen the sound transmission of structural borne noise, to cover noise barriers to make them more sustainable and weather resistant, to dampen the vibration of machines, and to reduce the structural vibration of buildings.

Task-based detectability in anatomical background in digital mammography, digital breast tomosynthesis and synthetic mammography

Objective. Determining the detectability of targets for the different imaging modalities in mammography in the presence of anatomical background noise is challenging. This work proposes a method to compare the image quality and detectability of targets in digital mammography (DM), digital breast tomosynthesis (DBT) and synthetic mammography. Approach. The low-frequency structured noise produced by a water phantom with acrylic spheres was used to simulate anatomical background noise for the different types of images. A method was developed to apply the non-prewhitening observer model with eye filter (NPWE) in these conditions. A homogeneous poly(methyl) methacrylate phantom with a 0.2 mm thick aluminium disc was used to calculate 2D in-plane modulation transfer function (MTF), noise power spectrum (NPS), noise equivalent quanta, and system detective quantum efficiency for 30, 50 and 70 mm thicknesses. The in-depth MTFs of DBT volumes were determined using a thin tungsten wire. The MTF, system NPS and anatomical NPS were used in the NPWE model to calculate the threshold gold thickness of the gold discs contained in the CDMAM phantom, which was taken as reference. Main results. The correspondence between the NPWE model and the CDMAM phantom (linear Pearson correlation 0.980) yielded a threshold detectability index that was used to determine the threshold diameter of spherical microcalcifications and masses. DBT imaging improved the detection of masses, which depended mostly on the reduction of anatomical background noise. Conversely, DM images yielded the best detection of microcalcification s. Significance. The method presented in this study was able to quantify image quality and object detectability for the different imaging modalities and levels of anatomical background noise.