Noise Rate Research Articles

Signal processing to reduce dark noise impact in precision timing

We introduce a technique to mitigate the effects of low frequency noise on precision timing.The example of Dark Count Noise Rate (DCR) in Silicon Photomultipliers (SiPMs) is emphasized.This technique exploits the correlation between time shifts onthe leading edge of a signal and the residual slope of the baseline (due to noise) which remainsafter baseline subtraction. In fast timing applications (such as for Time-of-flight particle ID) the signal arrival time is typically captured on the signal leading edge. The signal risetime is often fixed by the physics of the sensor and input circuit. Then accurate pulse timing can be achieved by correcting a leading edge threshold time (depending on a slope proportional to both the Amplitude and the risetime) to a “constant fraction” time. This compensation for time walk due to amplitude fluctuations breaks down once we introduce interference from low frequency noise on the leading edge. In this paper we demonstrate that an additional measurement of the slope at threshold can be used to correct for this noise jitter.

Boosting Facial Expression Recognition by A Semi-Supervised Progressive Teacher

In this paper, we aim to improve the performance of in-the-wild Facial Expression Recognition (FER) by exploiting semi-supervised learning. Large-scale labeled data and deep learning methods have greatly improved the performance of image recognition. However, the performance of FER is still not ideal due to the lack of training data and incorrect annotations (e.g., label noises). Among existing in-the-wild FER datasets, reliable ones contain insufficient data to train robust deep models while large-scale ones are annotated in lower quality. To address this problem, we propose a semi-supervised learning algorithm named Progressive Teacher (PT) to utilize reliable FER datasets as well as large-scale unlabeled expression images for effective training. On the one hand, PT introduces semi-supervised learning method to relieve the shortage of data in FER. On the other hand, it selects useful labeled training samples automatically and progressively to alleviate label noise. PT uses selected clean labeled data for computing the supervised classification loss and unlabeled data for unsupervised consistency loss. Experiments on widely-used databases RAF-DB and FERPlus validate the effectiveness of our method, which achieves state-of-the-art performance with accuracy of 89.57% on RAF-DB. Additionally, when the synthetic noise rate reaches even 30%, the performance of our PT algorithm only degrades by 4.37%.

Robust Wafer Classification With Imperfectly Labeled Data Based on Self-Boosting Co-Teaching

Wafer classification is a critical task for semiconductor manufacturing. Most conventional algorithms require a large-scale perfectly-labeled dataset to train accurate classifiers. In practice, it is usually difficult or even impossible to collect perfect labels without errors, and the classification accuracy in the presence of imperfectly labeled data would degrade. To facilitate robust wafer classification with noisy labels, we propose a novel self-boosting co-teaching (SB-CT) approach. Specifically, we iteratively correct the wrong labels by using the predictions of two classifiers that are jointly trained with noisily labeled data. To make the proposed method of practical utility, we develop a novel method to accurately estimate the noise rate, and adopt a probability scaling technique to further improve the classification accuracy. As demonstrated by the experimental results based on two industrial datasets, the proposed SB-CT approach achieves superior accuracy over other conventional methods.

Machine Learning based tool for CMS RPC currents quality monitoring

The muon system of the CERN Compact Muon Solenoid (CMS) experiment includes more than a thousand Resistive Plate Chambers (RPC). They are gaseous detectors operated in the hostile environment of the CMS underground cavern on the Large Hadron Collider where pp luminosities of up to 2×1034cm−2s−1 are routinely achieved. The CMS RPC system performance is constantly monitored and the detector is regularly maintained to ensure stable operation. The main monitorable characteristics are dark current, efficiency for muon detection, noise rate etc. Herein we describe an automated tool for CMS RPC current monitoring which uses Machine Learning techniques. We further elaborate on the dedicated generalized linear model proposed already and add autoencoder models for self-consistent predictions as well as hybrid models to allow for RPC current predictions in a distant future.

Development of a sealed MRPC with mylar spacers for high luminosity TOF systems

The high luminosity Time-of-Flight (TOF) systems call for timing detectors with both good time resolution and adequate rate capability, which have been fulfilled by the Multigap Resistive Plate Chamber (MRPC) assembled with low-resistive glass electrodes. However, recent tests and operations with high rate MRPCs have given evidences to the luminosity effects, such as the dark current rise and related higher noise rate, which may vitiate the stability in experimental operations. This work discusses the results from observations and tests, and infer the fishline spacers as a source of dark current and noise problems. Then, a sealed MRPC with mylar spacers is proposed, developed and tested. The prototype showed good luminosity tolerance during the X-ray test, and its performances of 95% efficiency and 80 ps time resolution have been validated in cosmic tests.

Neural-network-based level-1 trigger upgrade for the SuperCDMS experiment at SNOLAB

The extended physics program of the SuperCDMS SNOLAB dark matter search experiment aims to maximize the sensitivity to low-mass dark matter. To realize this, an upgrade of the existing level-1 trigger of the data acquisition system is proposed by making use of a recurrent neural network to be implemented on the trigger FPGA. This provides an improved amplitude estimator and signal-noise discriminator based on the combined information of filtered traces from individual detector channels. The architecture and configuration of this neural trigger are discussed in this article, and the improvements in key performance indicators such as the efficiency, resolution, and noise rate are quantified based on signal simulations and noise data. Based on the findings in this proof of concept, the trigger threshold is expected to be lowered by ∼ 22%.

Robust Point Cloud Segmentation with Noisy Annotations.

Point cloud segmentation is a fundamental task in 3D. Despite recent progress on point cloud segmentation with the power of deep networks, current learning methods based on the clean label assumptions may fail with noisy labels. Yet, class labels are often mislabeled at both instance-level and boundary-level in real-world datasets. In this work, we take the lead in solving the instance-level label noise by proposing a Point Noise-Adaptive Learning (PNAL) framework. Compared to noise-robust methods on image tasks, our framework is noise-rate blind, to cope with the spatially variant noise rate specific to point clouds. Specifically, we propose a point-wise confidence selection to obtain reliable labels from the historical predictions of each point. A cluster-wise label correction is proposed with a voting strategy to generate the best possible label by considering the neighbor correlations. To handle boundary-level label noise, we also propose a variant "PNAL-boundary" with a progressive boundary label cleaning strategy. Extensive experiments demonstrate its effectiveness on both synthetic and real-world noisy datasets. Even with 60% symmetric noise and high-level boundary noise, our framework significantly outperforms its baselines, and is comparable to the upper bound trained on completely clean data. Moreover, we cleaned the popular real-world dataset ScanNetV2 for rigorous experiment. Our code and data is available at https://github.com/pleaseconnectwifi/PNAL.

Light rail wheel/rail noise radiation model using a coupled FE-BE method

This paper presents the development of a coupled finite element - boundary element model of rolling noise from a resilient light rail wheel on several types of tracks. The vibro-acoustic model utilizes a 3D elastic solid representation for both wheel and rail which are excited by linearized Hertzian interface loads. Estimates of the radiated sound power, radiation efficiencies and wayside noise levels are obtained and then scaled using a parallel impedance framework and a roughness profile. Results are compared to recently measured vehicle passby noise and track decay rates at Sound Transit in Seattle Washington.

Performance enhancement of passive optical network with improved bandwidth utilisation and allocation

Abstract In the past decade optical network is growing very rapidly due to the low noise, less bit error rate as well as bandwidth utilization of the network. In this paper explored the bandwidth utilization with the improved gain with less noise and low bit error rate. Previously worked had been done on this area up to the 2nd stage of the optical networks. Previous worked explain the utilization of the bandwidth also explore the allocation of bandwidth in between the two different networks. They also find out the different optimized parameters like bit error rate, delay, gain, and low noise using different methods. But this paper explored up to the fifth stage of the optical network with the optimum parameters has been found out after the simulation. It also explains the bandwidth reuse to the next level of the optical network without any delay provided in between the two consecutive stages. In this explored the future prospective of the next generation optical network and also explain the different module of the stages.

Learning under concept drift and non-stationary noise: Introduction of the concept of persistence

Learning from noisy data is a challenging task especially when the system under consideration has a non-stationary nature. The source of the noise is often assumed to be stationary, however the severity or characteristics of noise may also be time-varying, which causes multiple sources of drift in the collected data. This study introduces a novel adaptive learning rate approach to improve learning when the observations from a non-stationary system is altered by an also non-stationary source of noise. As an example to this approach, we propose Persistence Aware Robust Learner (PeARL), an online learning method that utilizes a novel concept called persistence, which is a local noisiness estimation metric to measure the correspondence of a signal to discrete white noise. Making use of this metric, PeARL is able to adaptively adjust the learning rate for each observation during learning to reduce the effect of noise. With this level of control on the learning rate, noisy instances have less disruptive effect on the maintained estimate. We experimentally evaluate PeARL on (a) systematically generated synthetic data and (b) real-world data, including accelerometer readings collected from people (HASC2010corpus) and current measurements from electric motors collected within the scope of EU-funded research project iRel40. The experiments reveal a favorable region of noise rate, in which the proposed method achieves up to 40% reduction in mean absolute error (MAE).

Joint remote state preparation in multi-hop network under noisy environment

Joint remote state preparation is an important method to transmit quantum information with more senders and higher security. In this paper, we present a deterministic joint remote state preparation scheme in multi-hop network with two senders and N intermediate parties, using only projective measurements and recovery operations. We describe the scheme under the framework of density matrix to investigate the performance of the scheme in noisy environment. The relation of fidelity, noise rate and the number of intermediate nodes is given for three types of noise. It is revealed that the average fidelity attains its minimum when the noise rate is at the most uncertain point, decreases monotonically as the number of intermediate nodes increases. However, in some special cases, the average fidelity of the multi-hop scheme is greater than some existing one step joint remote state preparation scheme.

Comparative evaluation of three commercially available markerless depth sensors for close-range use in surgical simulation

PurposeMinimally invasive surgeries have restricted surgical ports, demanding a high skill level from the surgeon. Surgical simulation potentially reduces this steep learning curve and additionally provides quantitative feedback. Markerless depth sensors show great promise for quantification, but most such sensors are not designed for accurate reconstruction of complex anatomical forms in close-range.MethodsThis work compares three commercially available depth sensors, namely the Intel D405, D415, and the Stereolabs Zed-Mini in the range of 12–20 cm, for use in surgical simulation. Three environments are designed that closely mimic surgical simulation, comprising planar surfaces, rigid objects, and mitral valve models of silicone and realistic porcine tissue. The cameras are evaluated on Z-accuracy, temporal noise, fill rate, checker distance, point cloud comparisons, and visual inspection of surgical scenes, across several camera settings.ResultsThe Intel cameras show sub-mm accuracy in most static environments. The D415 fails in reconstructing valve models, while the Zed-Mini provides lesser temporal noise and higher fill rate. The D405 could reconstruct anatomical structures like the mitral valve leaflet and a ring prosthesis, but performs poorly for reflective surfaces like surgical tools and thin structures like sutures.ConclusionIf a high temporal resolution is needed and lower spatial resolution is acceptable, the Zed-Mini is the best choice, whereas the Intel D405 is the most suited for close-range applications. The D405 shows potential for applications like deformable registration of surfaces, but is not yet suitable for applications like real-time tool tracking or surgical skill assessment.

PO-03-007 CROSS VALIDATION OF A MACHINE LEARNING ALGORITHM DESIGNED TO IDENTIFY PULMONARY VEIN SIGNALS USING COMPETING TRAINING DATA

State of the art electrophysiology (EP) recording systems allow recording of both the surface electrocardiograms and intracardiac electrograms (EGMs) in an unfiltered fashion with improved quality, lower noise and higher sample rate. Legacy EP recording systems sample and store irreversibly pre-filtered EGMs, often susceptible to noise and interference. This study hypothesized that EGM quality impacts the performance of machine learning algorithms later trained on these data. This study compares the performance of a previously validated machine learning algorithm that classifies pulmonary vein potentials trained on simultaneous recorded legacy and novel EP recording system datasets. Data from pulmonary vein isolation procedures were simultaneously collected using a conventional EP recording system (EP-Tracer, CardioTek and LabSystem Pro, Boston Scientific) and a novel electrophysiology recording system (ECGenius System, CathVision, Denmark) connected in parallel. A machine learning classification algorithm, originally designed for legacy data, was adapted to use data from both the conventional and ECGenius EP recording systems. The performance of the algorithm was evaluated using k-fold validation without overlap in patient data between folds. Classified labels across all folds were summarized and compared to the common ground truth labels of the two datasets. In total, 1536 samples from 521 EGMs were extracted from 89 patients prospectively enrolled in the PVISION study. The trained algorithm using the ECGenius recorded EGMs resulted in a higher predictive value (91% vs 84%), sensitivity (93% vs 90%) and overall accuracy (91% vs 86%) compared to the algorithm trained with EGMs from the conventional dataset, resulting in superiority of the ECGenius System algorithm (p <0.05). Receiver operating characteristic (ROC) curves show superiority of the ECGenius-trained algorithm at all relevant operating points (Fig. 1). Performance of machine learning algorithms using intracardiac EP data are dependent on EGM quality and sample rate. Using high quality, unfiltered unipolar EGMs from a novel EP recording system resulted in superior machine learning algorithm performance compared to the use of EGMs recorded with legacy EP systems.

Effect of sound masking using child-friendly masker sounds on the perception of the dental drill noise in pediatric dental practice

Background: Dental anxiety is significantly influenced by the fear and annoyance brought on by the sound of a dental drill. As a result, the welcoming atmosphere at pediatric dental offices is likely to entice young children to come in for examination and treatment. Aim: The aim of this study was to compare the masking effect of dental drill noise with the sound maskers on the annoyance rate of pediatric dental patients. Materials and Methods: To evaluate the stimuli in the annoyance experiment, 40 children (20 boys and 20 girls) were enlisted. As masker sounds, dragonfly and whistle sounds were chosen, and the functioning dental drill noise was blended with the original sound level. The participants rated each audio stimulus signal delivered to them in a random order, with the level of annoyance indicated by a color scale corresponding to the intensity of their psychological sense. The mean comparison between the acoustic stimuli was made using the Friedman's test. Based on gender, the mean comparison between acoustic stimuli was made using the Mann–Whitney U-test. Results: The mean annoyance rate of dental drill noise dragonfly sound and whistle sound was 4.83, 1.15, and 3.82, respectively. Gender has a significant effect on masking effect in nonacoustic factor outcomes. Conclusion: The dragonfly sound had the best impact and the most significant reduction in annoyance compared to the sound of the drill. Typically, girls felt more at ease around the dragonfly sound than boys did.

Outlier galaxy images in the Dark Energy Survey and their identification with unsupervised machine learning

The Dark Energy Survey is able to collect image data of an extremely large number of extragalactic objects, and it can be reasonably assumed that many unusual objects of high scientific interest are hidden inside these data. Due to the extreme size of DES data, identifying these objects among many millions of other celestial objects is a challenging task. The problem of outlier detection is further magnified by the presence of noisy or saturated images. When the number of tested objects is extremely high, even a small rate of noise or false positives leads to a very large number of false detections, making an automatic system impractical. This study applies an automatic method for automatic detection of outlier objects in the first data release of the Dark Energy Survey. By using machine learning-based outlier detection, the algorithm is able to identify objects that are visually different from the majority of the other objects in the database. An important feature of the algorithm is that it allows to control the false-positive rate, and therefore can be used for practical outlier detection. The algorithm does not provide perfect accuracy in the detection of outlier objects, but it reduces the data substantially to allow practical outlier detection. For instance, the selection of the top 250 objects after applying the algorithm to more than 2⋅106 DES images provides a collection of uncommon galaxies. Such collection would have been extremely time-consuming to compile by using manual inspection of the data.

Multi-angle lensless ptychographic imaging via adaptive correction and the Nesterov method.

Lensless systems based on ptychographic imaging can simultaneously achieve a large field of view and high resolution while having the advantages of small size, portability, and low cost compared to traditional lensed imaging. However, lensless imaging systems are susceptible to environmental noise and have a lower resolution of individual images than lens-based imaging systems, which means that they require a longer time to obtain a good result. Therefore, in this paper, to improve the convergence rate and robustness of noise in lensless ptychographic imaging, we propose an adaptive correction method, in which we add an adaptive error term and noise correction term in lensless ptychographic algorithms to reach convergence faster and create a better suppression effect on both Gaussian noise and Poisson noise. The Wirtinger flow and the Nesterov algorithms are used in our method to reduce computational complexity and improve the convergence rate. We applied the method to phase reconstruction for lensless imaging and demonstrated the effectiveness of the method by simulation and experiment. The method can be easily applied to other ptychographic iterative algorithms.

A Pre-Pruning Quadtree Isolation Method with Changing Threshold for ICESat-2 Bathymetric Photon Extraction

The new generation of spaceborne laser altimeter, the Ice, Cloud, and Land Elevation Satellite-2 (ICESat-2), which can simultaneously generate laser reflections on the water surface and underwater, is a potential data source for exploring water depth in nearshore environments. To achieve this scientific goal, accurate bathymetric photon extraction is needed. This study proposed a pre-pruning quadtree isolation (PQI) method with changing threshold. Firstly, the pre-pruning step is introduced. Photons are transformed into different levels in the isolated quadtree structure according to spatial density. Then, the frequency histogram of photon elevation and isolated level (IL) is generated, the IL thresholds in different depth ranges are calculated by the Otsu method, and the bathymetric photons are extracted. The results in the Culebra archive show that this method achieved a 92.71% F1 score. Noise rate and water depth are the main factors affecting the extraction of sounding photons. When the photon density gradually increases from 2–4 pts/m to 6–8 pts/m, the F1 score of PQI decreases by no more than two percent. In different depth ranges, the extraction results of PQI are also better than those of comparison methods. Therefore, PQI can provide reliable theoretical support for nearshore areas lacking water depth data.

A late fusion deep neural network for robust speaker identification using raw waveforms and gammatone cepstral coefficients

Speaker identification aims at determining the speaker identity by analyzing his voice characteristics, and relies typically on statistical models or machine learning techniques. Frequency-domain features are by far the most used choice to encode the audio input in sound recognition. Recently, some studies have also analyzed the use of time-domain raw waveform (RW) with deep neural network (DNN) architectures. In this paper, we hypothesize that both time-domain and frequency-domain features can be used to increase the robustness of speaker identification task in adverse noisy and reverberation conditions, and we present a method based on a late fusion DNN using RWs and gammatone cepstral coefficients (GTCCs). We analyze the characteristics of RW and spectrum-based short-time features, reporting advantages and limitations, and we show that the joint use can increase the identification accuracy. The proposed late fusion DNN model consists of two independent DNN branches made primarily by convolutional neural networks (CNN) and fully connected neural networks (NN) layers. The two DNN branches have as input short-time RW audio fragments and GTCCs, respectively. The late fusion is computed on the predicted scores of the DNN branches. Since the method is based on short segments, it has the advantage of being independent from the size of the input audio signal, and the identification task can be computed by summing the predicted scores over several short-time frames. Analysis of speaker identification performance computed with simulations show that the late fusion DNN model improves the accuracy rate in adverse noise and reverberation conditions in comparison to the RW, the GTCC, and the mel-frequency cepstral coefficients (MFCCs) features. Experiments with real-world speech datasets confirm the efficiency of the proposed method, especially with small-size audio samples.

Hyper-Dimensional Modulation for Robust Short Packets in Massive Machine-Type Communications

In this paper, we introduce Hyper-Dimensional Modulation (HDM) for massive machine-type communications (mMTC). HDM enables robust communication of short packets by spreading information bits across many elements in a hyper-dimensional vector and superimposing a set of such non-orthogonal vectors. The proposed CRC-aided K-best decoding algorithm for HDM can achieve a very low packet error rate (PER) in additive white Gaussian noise (AWGN) channels for short packets. Furthermore, extended decoding algorithms are proposed to combat overwhelming interference in an mMTC network. Comprehensive simulation and real-world experiment results show that HDM outperforms sparse superposition codes in AWGN channels and state-of-the-art short codes such as polar and tail-biting convolutional codes in interference-heavy channels for short packet transmissions.

Accent rating of noise- and tone-vocoded foreign-accented speech

There are abundant acoustic-phonetic cues in speech signals for listeners to encode talker identity. However, speech signals in the real world are always less optimal due to various adverse listening sources. Vocoded speech is one type of simplified signal that has less spectral and/or temporal information in comparison to normal speech. The purpose of this study is to examine whether and how listeners’ judgment of talker accent is affected by noise and tone vocoding. Twelve Mandarin-accented English speakers with varying degree of accentedness and two native English speakers were recorded reading “The Rainbow Passage.” The recorded speech samples from each talker were segmented into small sections that were randomly selected for noise- or tone-excited vocoder processing into 1, 2, 4, 8, and 16 channels. The vocoded and unprocessed speech samples were randomly presented to a group of normal-hearing, monolingual English listeners for accent rating. The listeners judged the degree of talker accent on a 9-point Likert scale with “1” representing no accent and “9” representing extremely strong accent. The data are still in the process of being collected and analyzed. Results and implications of the present study will be discussed.