Noise Classification Research Articles

DOL3: Distilled OpenL3 audio embeddings for lightweight audio classification

Deep audio representations, also known as embeddings, recently became a popular alternative to conventional features like spectrograms for a wide range of audio classification tasks because of their domain-agnostic character and reduced training costs. Still, the usage is often limited to rather computationally intensive system due to the nature of their extraction from large networks. This paper aims to minimize the computational costs of embedding extraction by distilling the knowledge of the OpenL3 audio network to a smaller student network. Results show that the student network maintains comparable performance as the teacher network on various music and ambient noise classification tasks, while reducing the network size by over 90\% and the computational load by five times.

Classification and detection of noise in IoT based MQ gas sensors using artificial neural network

Abstract IoT-based Metal Oxide Semiconductor (MOS) gas sensors have potential applications in health, industrial, and agriculture sectors. MQ gas sensors are easy to use, have a large detection range, high sensitivity, and efficiency, and can be interfaced with Arduino for effective use. However, like other any other sensors, the MQ gas sensors are also not able to escape the effect of noise that affects the sensitivity and selectivity causing the unwanted wrong classification and identification of gases, etc finally leading to a wrong justification of results and data. Hence, the study of noise and its removal is very crucial for greater accuracy in its analysis. The main aim of our work is to check whether one can classify different types of noises or noisy signals accurately using even simple ANNs so that after testing the different filtration techniques for the signals we can obtain an independent system specifically for the MQ gas sensors, that can classify as well as filter the noisy signals based on the category to which they are classified. The best and most effective noise filtration method is then obtained. During the study Narrow Neural Network model and Medium Neural Network demonstrated high accuracy in validation and testing, with ROC curves indicating their efficiency and effectiveness.

Exploring the efficacy of GRU model in classifying the signal to noise ratio of microgrid model

Microgrids are small-scale energy system that supplies power to homes, businesses, and industries. Microgrids can be considered as a trending technology in energy fields due to their power to supply reliable and sustainable energy. Microgrids have a mode called the island, in this mode, microgrids are disconnected from the major grid and keep providing energy in the situation of an energy outage. Therefore, they help the main grid during peak energy demand times. The microgrids can be connected to the network, which is called networked microgrids. It is possible to have flexible energy resources by using their enhanced energy management systems. However, connection microgrid systems to the communication network introduces various challenges, including increased in systems complicity and noise interference. Integrating network communication into a microgrid system causes the system to be susceptible to noise, potentially disrupting the critical control signals that ensure smooth operation. Therefore, there is a need for predicting noise caused by communication network to ensure the operation stability of microgrids. In addition, there is a need for a simulation model that includes communication network and can generate noise to simulate real scenarios. This paper proposes a classifying model named Noise Classification Simulation Model (NCSM) that exploits the potential of deep learning to predict noise levels by classifying the values of signal-to-noise ratio (SNR) in real-time network traffic of microgrid system. This is accomplished by initially applying Gaussian white noise into the data that is generated by microgrid model. Then, the data has noise and data without noise is transmitted through serial communication to simulate real world scenario. At the end, a Gated Recurrent Unit (GRU) model is implemented to predict SNR values for the network traffic data. Our findings show that the proposed model produced promising results in predicting noise. In addition, the classification performance of the proposed model is compared with well-known machine learning models and according to the experimental results, our proposed model has noticeable performance, which achieved 99.96% classification accuracy.

Edge-Based Data Sensing and Processing Platform for Urban Noise Classification

Edge-Based Data Sensing and Processing Platform for Urban Noise Classification

A robust one-stage detector for SAR ship detection with sequential three-way decisions and multi-granularity

Synthetic Aperture Radar (SAR) images are widely used in ship detection because of their all-weather and all-day imaging characteristics. However, there are two challenges for SAR ship detection. One is coherent speckle noise, causing ship confusion with similar objects and raising false alarms. The other is multi-scale ship detection, particularly in small ships, which suffers from insufficient accuracy. To address these challenges, this paper proposes a robust one-stage detector, S3MDet, for SAR ship detection with sequential three-way decisions (S3WDs) and multi-granularity. First, to effectively eliminate the interference of coherent speckle noise, a noise classification and denoising module (NCDM) S3WD-based is designed. This module can accurately classify the noise level of the image and only perform denoising on the images identified as noisy, avoiding unnecessary operations on noise-free images. Then, to solve the problem of multi-scale ship detection, a multi-granularity group attention module (MGAM) is designed to obtain a richer representation of multi-granularity features. This module adopts a multi-granularity group convolution structure and channel-wise attention weights to efficiently extract ship features of different scales from SAR images. Extensive experiments on four SAR ship datasets, including SAR-Ship-Dataset, HRSID, SSDD, and LS-SSDD-v1.0, validate the robustness of S3MDet, demonstrating that it achieves state-of-the-art performance.

Deep Learning Based Underwater Acoustic Target Recognition: Introduce a Recent Temporal 2D Modeling Method.

In recent years, the application of deep learning models for underwater target recognition has become a popular trend. Most of these are pure 1D models used for processing time-domain signals or pure 2D models used for processing time-frequency spectra. In this paper, a recent temporal 2D modeling method is introduced into the construction of ship radiation noise classification models, combining 1D and 2D. This method is based on the periodic characteristics of time-domain signals, shaping them into 2D signals and discovering long-term correlations between sampling points through 2D convolution to compensate for the limitations of 1D convolution. Integrating this method with the current state-of-the-art model structure and using samples from the Deepship database for network training and testing, it was found that this method could further improve the accuracy (0.9%) and reduce the parameter count (30%), providing a new option for model construction and optimization. Meanwhile, the effectiveness of training models using time-domain signals or time-frequency representations has been compared, finding that the model based on time-domain signals is more sensitive and has a smaller storage footprint (reduced to 30%), whereas the model based on time-frequency representation can achieve higher accuracy (1-2%).

Machine learning for analysis of wind farm noise

Two challenges associated with analyzing acoustical data from wind farms are: 1) separating turbine sounds from environmental sounds and 2) classifying acoustical samples into different types of wind turbine noise based on acoustic characteristics. Machine-learning methods for classifying general environmental sounds have been developed using large human classified databases (e.g., YAMNet), but only a few studies have targeted classification of wind farm noise (WFN) specifically. Techniques for classifying wind farm noise have focused on identification of amplitude modulation (AM) using both traditional methods such as low frequency peak prominence (IOA method) and machine learning methods using both targeted AM acoustics features and more general deep acoustic features. To address these two challenges, we are developing a multi-echelon machine learning framework to identify and classify noise from wind farms using publicly available windfarm data and open-source software. The first echelon provides an automated method for identifying WFN samples that are free of environmental sounds. The second echelon uses machine learning to classify these wind farm noises according to the degree of AM, prominent tones, and other factors that might contribute to the human response and are incorporated in the metrics used or potentially used to assess compliance with regulations.

Speech Enhancement Using Joint DNN-NMF Model Learned with Multi-Objective Frequency Differential Spectrum Loss Function

We propose a multi-objective joint model of non-negative matrix factorization (NMF) and deep neural network (DNN) with a new loss function for speech enhancement. The proposed loss function (LMOFD) is a weighted combination of a frequency differential spectrum mean squared error (MSE)-based loss function (LFD) and a multi-objective MSE loss function LMO. The conventional MSE loss function computes the discrepancy between the estimated speech and clean speech across all frequencies, disregarding the process of changing amplitude in the frequency domain which contains valuable information. The differential spectrum representation retains spectral peaks that carry important information. Using this representation helps to ensure that this information in the speech signal is reserved. Also, on the other hand, noise spectra typically have a flat shape and as the differential operation makes the flat spectral partly close to zero, the differential spectrum is resistant to noises with smooth structures. Thus, we propose using a frequency-differentiated loss function that considers the magnitude spectrum differentiations between the neighboring frequency bins in each time frame. This approach maintains the spectrum variations of the objective signal in the frequency domain, which can effectively reduce the noise deterioration effects. The multi-objective MSE term LMO is a combined two-loss function related to the NMF coefficients which are the intermediate output targets, and the original spectral signals as the actual output targets. The use of encoded NMF coefficients as low-dimensional structural features for DNN serves as prior knowledge and helps the learning process. LMO is used beside LFD to take advantage of both the properties of the original and the differential spectrum in the training loss function. Moreover, a DNN-based noise classification and fusion strategy (NCF) is proposed to exploit a discriminative model for noise reduction. The experiments reveal the improvements of the proposed approach compared to the previous methods.

Optimized systolic array filters with noise classification for extracting FECG

To preserve the lives of both mother and foetus during the early phases of childbirth, heart abnormality diagnosis is essential. In remote places where understanding of maternal care is limited, the death rate caused by carelessness or a failure to detect abnormalities is still a problem. The signal will contain various external or internal noise sources to isolate the key components of the foetal heart rate from the mother's belly during labour. The likelihood that the genuine signal would be misinterpreted and result in a false report will increase in the presence of such noise sources. Although there are many software programs available for extracting the QRS features from foetal ECG signals, it is unavoidable that specialized hardware is required for a significant reduction in both area and power. This paper's main goal is to extract the QRS complex using LDA, then improve Social Spider classifier performance using the suggested TAODV as a distance metric calculator, and then compare against existing methods to discover sounds that are distorting the normal heart rate. Systolic array filter with suggested Glitch Avoidance Circuit employing MUX is simulated using Cadence Virtuoso in 65nm technology to remove noise from the observed QRS complex. Over 100 records with the necessary examples from MIT-BIH Arrhythmia were used in the simulations, and it was discovered that MATLAB 2010b was used to adopt a unique technique for classifying noise. The suggested TAODV-based SSA classifier's accuracy is 96.8%, whereas the accuracy of a filter with a glitch avoidance circuit is 96.13%. The primary benefit of these strategies comprises cutting-edge hardware and computational solutions.

A Lightweight Network Based on Multi-Scale Asymmetric Convolutional Neural Networks with Attention Mechanism for Ship-Radiated Noise Classification

Ship-radiated noise classification is critical in ocean acoustics. Recently, the feature extraction method combined with time–frequency spectrograms and convolutional neural networks (CNNs) has effectively described the differences between various underwater targets. However, many existing CNNs are challenging to apply to embedded devices because of their high computational costs. This paper introduces a lightweight network based on multi-scale asymmetric CNNs with an attention mechanism (MA-CNN-A) for ship-radiated noise classification. Specifically, according to the multi-resolution analysis relying on the relationship between multi-scale convolution kernels and feature maps, MA-CNN-A can autonomously extract more fine-grained multi-scale features from the time–frequency domain. Meanwhile, the MA-CNN-A maintains its light weight by employing asymmetric convolutions to balance accuracy and efficiency. The number of parameters introduced by the attention mechanism only accounts for 0.02‰ of the model parameters. Experiments on the DeepShip dataset demonstrate that the MA-CNN-A outperforms some state-of-the-art networks with a recognition accuracy of 98.2% and significantly decreases the parameters. Compared with the CNN based on three-scale square convolutions, our method has a 68.1% reduction in parameters with improved recognition accuracy. The results of ablation explorations prove that the improvements benefit from asymmetric convolution, multi-scale block, and attention mechanism. Additionally, MA-CNN-A shows a robust performance against various interferences.

Underwater Targets Radiated Noise Classification Based on Enhanced Images and Convolutional Neural Networks

Underwater Targets Radiated Noise Classification Based on Enhanced Images and Convolutional Neural Networks

Disentangled Prototypical Convolutional Network for Few-Shot Learning in In-Vehicle Noise Classification

Disentangled Prototypical Convolutional Network for Few-Shot Learning in In-Vehicle Noise Classification

Variable-Step Multiscale Katz Fractal Dimension: A New Nonlinear Dynamic Metric for Ship-Radiated Noise Analysis

The Katz fractal dimension (KFD) is an effective nonlinear dynamic metric that characterizes the complexity of time series by calculating the distance between two consecutive points and has seen widespread applications across numerous fields. However, KFD is limited to depicting the complexity of information from a single scale and ignores the information buried under different scales. To tackle this limitation, we proposed the variable-step multiscale KFD (VSMKFD) by introducing a variable-step multiscale process in KFD. The proposed VSMKFD overcomes the disadvantage that the traditional coarse-grained process will shorten the length of the time series by varying the step size to obtain more sub-series, thus fully reflecting the complexity of information. Three simulated experimental results show that the VSMKFD is the most sensitive to the frequency changes of a chirp signal and has the best classification effect on noise signals and chaotic signals. Moreover, the VSMKFD outperforms five other commonly used nonlinear dynamic metrics for ship-radiated noise classification from two different databases: the National Park Service and DeepShip.

USING OPEN SOURCE NEURAL NETWORKS FOR NOISE CLASSIFICATION IN PRODUCTION AS NOVICE TO MACHINE LEARNING: SETTING FOOT IN THE WORLD OF MACHINE LEARNING AS COMPANY WITHOUT PRIOR KNOWLEDGE

An overwhelming amount of data processing is being taken over by machine learning. These algorithms are either open source academic works or closed off proprietary products. Especially the field of noise classification for noise monitoring is of great interest to both profit oriented noise management industry and academic psycho-acoustic research. In order to soften up this dichotomy of "open" and "closed", it shall be investigated how well a company without any prior machine learning knowledge can set foot in the world of machine learning using only open access methods to create a usable product. To combine the spirits of corporate interest and open source, this paper aims to provide the findings of audio and acoustics measurement company NTi-Audio AG on the quest for a universal noise classification system for their noise monitoring solutions. The road taken shows how an open source pre-trained convolutional neural network can be combined with post-processing to achieve noise classification all around the world, while also showing the dead-ends developers without any training in machine learning run into.

A ship-radiated noise classification method based on domain knowledge embedding and attention mechanism

Ship classification based on machine learning (ML) has proven to be a significant underwater acoustic research direction. One of the critical challenges rests with how to embed domain signal knowledge into ML models to obtain suitable features that highly correlate with the classification and create better predictors. In this paper, a novel ML-based ship classification model, Hierarchical Underwater Acoustic Transformer (HUAT), is proposed to improve the classification performance. Firstly, the Detection of Envelope Modulation on Noise (DEMON) spectra of ship-radiated noise signals are estimated by cyclostationary analysis. The motivation for using a DEMON-based preprocessing scheme is that valuable propeller information can be revealed by exploiting the second-order cyclostationarity of ship-radiated noise signals. Secondly, the useful features of DEMON spectra are enhanced using a multi-head self-attention module, and the potential features of the Mel spectrograms are extracted employing a Convolutional Neural Network (CNN) module. The two kinds of features are fused to provide ship classification patterns. The challenge of feature learning in the deep classification model is reduced by leveraging domain-related classification knowledge. Finally, the Swin Transformer, based on shifted window self-attention mechanism, is used to learn high-level feature representations and conduct ship classification. Experimental results show that the HUAT model achieves excellent classification performance on ship-radiated noise datasets, ShipsEar and DeepShip. And its classification efficiency is better than the model based on traditional Transformer architecture. In addition, the proposed method provides technical support for the underwater intelligent system capable of automatically sensing sailing vessels and recognizing vessel types.

Efficient ship noise classification with positive incentive noise and fused features using a simple convolutional network

Ship noise analysis is a critical area of research in hydroacoustic remote sensing due to its practical implications in identifying ship direction, type, and even specific ship identities. However, the limited availability of data poses challenges in developing accurate ship noise classification models. Previous studies have mainly focused on small-sample learning approaches, resulting in complex network structures. Nonetheless, underwater robots often have limited computing power, making it essential to develop simpler recognition networks. In this paper, we address the issue of data scarcity by introducing positive incentive noise. We propose a CNN-based hydroacoustic signal recognition method that achieves comparable or superior performance to previous studies, using a simple network structure as a back-end decision system. We describe the feature extraction process using a dataset with added noise and compare the performance of various features. Additionally, we compare our proposed method with previous studies. Experimental results demonstrate that simple neural networks can achieve high performance and excellent generalizability without the need for complex network structures like adversarial learning models.

Active Bayesian Deep Learning With Vector Sensor for Passive Sonar Sensing of the Ocean

Multiple studies have explored and reported on using passive sonar scalar pressure fields for both biological and man-made target classification. Here, inclusion of the vector field is considered and evaluated as an input feature in the classification of ship noises. We find that the inclusion of vector field information can significantly improve the understanding of the classification model performance. Based on an analysis of the epistemic and aleatoric uncertainties of probabilistic classification models, using vector field information as additional inputs leads to an <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$F_{1}$</tex-math></inline-formula> score of 0.8 in classification, comparable to scalar pressure field inputs only, but with 39% lower epistemic uncertainty in favor of utilizing the vector field component. Moreover, we verify that a similar conclusion applies to the active learning scenario, where we demonstrate the ability to utilize vector field information and epistemic uncertainty estimates to train a model that achieves an <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$F_{1}$</tex-math></inline-formula> score of 0.8 while using only 23% of the overall data. The ability to decompose uncertainty into an aleatoric and epistemic component leads to additional model explainability.

Underwater Noise Modeling and Its Application in Noise Classification with Small-Sized Samples

Underwater noise classification is of great significance for identifying ships as well as other vehicles. Moreover, it is helpful in ensuring a marine habitat-friendly, noise-free ocean environment. But a challenge we are facing is the small-sized underwater noise samples. Because noise is influenced by multiple sources, it is often difficult to determine and label which source or which two sources are dominant. At present, research to solve the problem is focused on noise image processing or advanced computer technology without starting with the noise generation mechanism and modeling. Here, a typical underwater noise generation model (UNGM) is established to augment noise samples. It is established by generating noise with certain kurtosis according to the spectral and statistical characteristics of the actual noise and filter design. In addition, an underwater noise classification model is developed based on UNGM and convolutional neural networks (CNN). Then the UNGM-CNN-based model is used to classify nine types of typical underwater noise, with either the 1/3 octave noise spectrum level (NSL) or power spectral density (PSD) as the input features. The results show that it is effective in improving classification accuracy. Specifically, it increases the classification accuracy by 1.59%, from 98.27% to 99.86%, and by 2.44%, from 97.45% to 99.89%, when the NSL and PSD are used as the input features, respectively. Additionally, the UNGM-CNN-based method appreciably improves macro-precision and macro-recall by approximately 0.87% and 0.83%, respectively, compared to the CNN-based method. These results demonstrate the effectiveness of the UNGM established in noise classification with small-sized samples.

Improved pitch shifting data augmentation for ship-radiated noise classification

The limited amount of ship-radiated noise data causes machine learning models to be prone to overfitting in training, and data augmentation methods could improve model generalization performance. The frequency stability of harmonic line spectra in ship-radiated noise leads to a lack of sample diversity generated by the pitch shifting method, which is overcome by the proposed improved method. Nine classification algorithms combining three time-frequency features and three classifiers are implemented and evaluated on the DeepShip and the ShipsEar datasets. The average accuracy increased by 1.67% on DeepShip and 2.25% on ShipsEar, using improved pitch shifting and time stretching augmentation methods. The constant-Q transformation-convolutional neural networks (CQT-CNN) performs best among these nine algorithms. Its accuracy improved from 68.33% to 74.08% on the DeepShip, and the F1 score of it improved from 57.92% to 61.45% on the ShipsEar. Data augmentation improves classification performance for each class of ships in different ways, suggesting that augmentation specific to the class and state of the ship would improve classification performance further.

Low-Power FPGA Realization of Lightweight Active Noise Cancellation with CNN Noise Classification

Active noise cancellation (ANC) is the most important function in an audio device because it removes unwanted ambient noise. As many audio devices are increasingly equipped with digital signal processing (DSP) circuits, the need for low-power and high-performance processors has arisen because of hardware resource restrictions. Low-power design is essential because wireless audio devices have limited batteries. Noise cancellers process the noise in real time, but they have a short secondary path delay in conventional least mean square (LMS) algorithms, which makes implementing high-quality ANC difficult. To solve these problems, we propose a fixed-filter noise cancelling system with a convolutional neural network (CNN) classification algorithm to accommodate short secondary path delay and reduce the noise ratio. The signal-to-noise ratio (SNR) improved by 2.3 dB with CNN noise cancellation compared to the adaptive LMS algorithm. A frequency-domain noise classification and coefficient selection algorithm is introduced to cancel the noise for time-varying systems. Additionally, our proposed ANC architecture includes an even–odd buffer that efficiently computes the fast Fourier transform (FFT) and overlap-save (OLS) convolution. The simulation results demonstrate that the proposed even–odd buffer reduces processing time by 20.3% and dynamic power consumption by 53% compared to the single buffer.