Higher Order Spectral Features Research Articles

Machine learning approach for detecting Covid-19 from speech signal using Mel frequency magnitude coefficient.

The Covid-19 pandemic is one of the most significant global health concerns that have emerged in this decade. Intelligent healthcare technology and techniques based on speech signal and artificial intelligence make it feasible to provide a faster and more efficient timely detection of Covid-19. The main objective of our study is to design speech signal-based noninvasive, low-cost, remote diagnosis of Covid-19. In this study, we have developed system to detect Covid-19 from speech signal using Mel frequency magnitude coefficients (MFMC) and machine learning techniques. In order to capture higher-order spectral features, the spectrum is divided into a larger number of subbands with narrower bandwidths as MFMC, which leads to better frequency resolution and less overall noise. As a consequence of an improvement in frequency resolution as well as a decrease in the quantity of noise that is included with the extraction of MFMC, the higher-order MFMCs are able to identify Covid-19 from speech signals with an increased level of accuracy. The procedures for machine learning are often less complicated than those for deep learning, and they may commonly be carried out on regular computers. However, deep learning systems need extensive computing power and data storage. Twelve, twenty-four, thirty, and forty spectral coefficients are obtained using MFMC in our study, and from these coefficients, performance is accessed using machine learning classifiers, such as random forests and K-nearest neighbor (KNN); however, KNN has performed better than the other model with having AUC score of 0.80.

Fusing features of speech for depression classification based on higher-order spectral analysis

Approximately 300 million people worldwide suffer from depression, and more than 60% of psychiatric patients do not have access to mental health services due to the shortage of psychiatrists and the high costs associated with clinical diagnosis and treatment. Correct and efficient diagnosis of depression can help overcome these straits. Automatic detection of depressive symptoms can help improve the accuracy and availability of diagnosis. In this paper, a fusion feature for Bispectral Features and Bicoherent Features by using higher-order spectral analysis. Experiments were performed on the Depression Sub-Challenge Dataset of the Audio/Visual Emotion Challenge 2017. The fusion feature fuses higher-order spectral features and traditional speech features with classification weights greater than 100 extracted by using A Collaborative Voice Analysis Repository. The support vector machine and k-nearest neighbor classification algorithms were used as the traditional machine learning models, and the convolutional neural network was used as the deep learning model to verify the proposed features. The experimental results show that under the support vector machine algorithm, the accuracies of extraction of speech-related features by using a collaborative voice analysis repository, The higher-order spectral analysis, and their fusion features were 63.15%, 68.42%, and 73.68%, respectively. Under the k-nearest neighbor classification algorithms model algorithm, the corresponding accuracies were 68.18%, 72.73%, and 77.27%, respectively. For the convolutional neural network model, the corresponding accuracies were 70%, 77%, and 85%, respectively. The results demonstrate that the fusion feature recognition accuracy is high and can be employed to improve the accuracy of depression identification by using traditional machine learning and deep learning models.

HOS-FingerCode: Bispectral invariants based contactless multi-finger recognition system using ridge orientation and feature fusion

Fingerprints are a commonly used biometric for civil and forensic applications and the majority of them are captured by contact sensors. More recently, there is a trend toward the use of high resolution digital and video cameras to acquire contactless images. However, the processing of such images is challenging because of intra-class variations such as pose, rotation, translation and scaling due to hand orientation and deformation. They can, however, facilitate the use of full finger or multiple fingers or hand information rather than the fingertip alone. Although some recent research has investigated representation of full finger and hand information under relaxed conditions, current state-of-the-art methods are unable to address transformation issues without manual intervention. To overcome the issues of relaxed constrained nature, geometrical variations and invariant feature encoding regardless of finger orientation, we have used higher order spectral features (HOS-FingerCode) from ridge patterns. We employ an automated method to detect key points and constructs a graph of key-lines. Ridge orientation profiles along with the selected key-lines provide a 1D signal from which bispectral invariant features are extracted. In addition, two fusion schemes are proposed where fusion from multiple key lines combinations and feature fusion from multiple key line configurations are investigated to extract ridge pattern profile for a better performance. The features are pooled from multiple fingers by concatenation or compact bilinear pooling to yield the HOS-FingerCode for multi-finger biometrics. For a single-finger system using 3369 high resolution index and middle finger images, the algorithm better performs when different key lines configurations are fused. For geometrical variations of the images, the system is most impacted by pose than rotation and scale changes of the images, and is tolerant to such variations. For the multi-finger system using 3111 images from both fingers, the algorithm has achieved 93.8% TPR and 97.95% classification accuracy at a setting of 2% FPR. For the feature fusion scheme proposed for multi-finger system using concatenation and compact bilinear pooling, concatenation of features performs better than the compact bilinear pooling. Since multi-finger recognition system performed better than the single finger biometrics system, it can help to diminish the high traffic scenario in busy premises and will facilitate the soft identity verification.

Cloud based ensemble machine learning approach for smart detection of epileptic seizures using higher order spectral analysis.

The present paper proposes a smart framework for detection of epileptic seizures using the concepts of IoT technologies, cloud computing and machine learning. This framework processes the acquired scalp EEG signals by Fast Walsh Hadamard transform. Then, the transformed frequency-domain signals are examined using higher-order spectral analysis to extract amplitude and entropy-based statistical features. The extracted features have been selected by means of correlation-based feature selection algorithm to achieve more real-time classification with reduced complexity and delay. Finally, the samples containing selected features have been fed to ensemble machine learning techniques for classification into several classes of EEG states, viz. normal, interictal and ictal. The employed techniques include Dagging, Bagging, Stacking, MultiBoost AB and AdaBoost M1 algorithms in integration with C4.5 decision tree algorithm as the base classifier. The results of the ensemble techniques are also compared with standalone C4.5 decision tree and SVM algorithms. The performance analysis through simulation results reveals that the ensemble of AdaBoost M1 and C4.5 decision tree algorithms with higher-order spectral features is an adequate technique for automated detection of epileptic seizures in real-time. This technique achieves 100% classification accuracy, sensitivity and specificity values with optimally small classification time.

Fault Classification of Nonlinear Small Sample Data through Feature Sub-Space Neighbor Vote

The fault classification of a small sample of high dimension is challenging, especially for a nonlinear and non-Gaussian manufacturing process. In this paper, a similarity-based feature selection and sub-space neighbor vote method is proposed to solve this problem. To capture the dynamics, nonlinearity, and non-Gaussianity in the irregular time series data, high order spectral features, and fractal dimension features are extracted, selected, and stacked in a regular matrix. To address the problem of a small sample, all labeled fault data are used for similarity decisions for a specific fault type. The distances between the new data and all fault types are calculated in their feature subspaces. The new data are classified to the nearest fault type by majority probability voting of the distances. Meanwhile, the selected features, from respective measured variables, indicate the cause of the fault. The proposed method is evaluated on a publicly available benchmark of a real semiconductor etching dataset. It is demonstrated that by using the high order spectral features and fractal dimensionality features, the proposed method can achieve more than 84% fault recognition accuracy. The resulting feature subspace can be used to match any new fault data to the fingerprint feature subspace of each fault type, and hence can pinpoint the root cause of a fault in a manufacturing process.

Research on Fingerprint Identification of Wireless Devices Based on Information Fusion

With the rapid development of wireless devices in recent years, the hardware tolerance of wireless devices has gradually become narrowed. Traditional radio frequency fingerprint(RF fingerprint) recognition methods are usually used based on single signal features, which will fail to characterize the subtle differences of wireless devices. Therefore, aiming at the shortcomings of traditional radio frequency fingerprint recognition methods, a multi-segment fusion recognition model is proposed based on D-S evidence theory. The fusion features of time-domain RF-DNA and high-order spectral features are used to obtain more accurate radio frequency fingerprint features. Simulation experiments show that the fusion method can significantly improve the recognition performance of traditional fingerprint recognition methods. When the SNR is higher than 5 dB, with the increasing number of signal fusion segment, the recognition rate of the proposed model will be higher than 99%, which prove that it has a better performance and can be used in practice.

Analysis of efficient Optimization Algorithm for Chaotic Information Nodes in Wireless Networks

Aiming improve the ability of fault detection and diagnosis of wireless network communication, the network chaotic information nodes are detected accurately and efficiently. An efficient detection method for chaotic information nodes in wireless networks is proposed based on quantization tracking and identification of sensor information. Firstly, the node connectivity graph model of wireless network communication is constructed, the link transmission channel of wireless network communication is analyzed, and the topology structure of distributed node is designed. Then the distributed sensor detection technique is used to detect the node output link equalization of wireless network communication. The higher-order spectral features of the output bit sequence stream of wireless network communication channel are extracted, and the anomaly of the characteristic quantity is extracted according to the characteristic quantity. The sensor information quantization tracking and identification technology is used to detect the chaotic information nodes in wireless network. Finally, the simulation test of high efficiency optimization detection of chaotic information nodes in wireless network communication is carried out. The results show that the accuracy of efficient optimization detection for chaotic information nodes is high by using this method. The quality of wireless network communication is improved because of the good real-time performance of network interrupt detection and fault diagnosis.

Remote Sensing for Vital Information Based on Spectral-Domain Harmonic Signatures

In this paper, we propose an efficient and robust non-contact heart rate estimation algorithm. The key idea is to recover the fundamental heartbeat frequency from its higher-order spectral features. Our contributions are threefold: 1) carefully review and examine possible approaches in vital signs detection with ultra-wide band (UWB) impulse radar; 2) numerically and experimentally show an important result that, in the spectral domain, the fundamental heartbeat signal is respiration-interference-limited, whereas its higher-order harmonics are noise-limited; and 3) implement an adaptive heart rate monitoring algorithm based on the proposed theory, which is feasible with continuous monitoring. To justify the proposed theory, we perform a spectral analysis of the harmonics of vital signs signal. We validate the proposed algorithm using a controlled vital sign simulator. We experimentally demonstrate the effectiveness of the harmonics-based heart rate estimation algorithm and compare it against existing methods. For completeness, we also provide a limitation analysis of the proposed algorithm.

Higher-order spectral analysis of spontaneous speech signals in Alzheimer's disease.

An early and accurate diagnosis of Alzheimer's disease (AD) has been progressively attracting more attention in recent years. One of the main problems of AD is the loss of language skills. This paper presents a computational framework for classifying AD patients compared to healthy control subjects using information from spontaneous speech signals. Spontaneous speech data are obtained from 30 AD patients and 30 healthy controls. Because of the nonlinear and dynamic nature of speech signals, higher order spectral features (specifically bispectrum) were used for analysis. Four classifiers (k-Nearest Neighbor, Support Vector Machine, Naïve Bayes and Decision tree) were used to classify subjects into three different levels of AD and healthy group based on their performance in terms of the HOS-based features. Ten-fold cross-validation method was used to test the reliability of the classifier results. The results showed that the proposed method had a good potential in AD diagnosis. The proposed method was also able to diagnose the earliest stage of AD with high accuracy. The method has the great advantage of being non-invasive, cost-effective, and associated with no side effects. Therefore, the proposed method can be a spontaneous speech directed test for pre-clinical evaluation of AD diagnosis.

A new hybrid PSO assisted biogeography-based optimization for emotion and stress recognition from speech signal

Speech signals and glottal signals convey speakers’ emotional state along with linguistic information. To recognize speakers’ emotions and respond to it expressively is very much important for human-machine interaction. To develop a subject independent speech emotion/stress recognition system, by identifying speaker's emotion from their voices, features from OpenSmile toolbox, higher order spectral features and feature selection algorithm, is proposed in this work. Feature selection plays an important role in overcoming the challenge of dimensionality in several applications. This paper proposes a new particle swarm optimization assisted Biogeography-based algorithm for feature selection. The simulations were conducted using Berlin Emotional Speech Database (BES), Surrey Audio-Visual Expressed Emotion Database (SAVEE), Speech under Simulated and Actual Stress (SUSAS) and also validated using eight benchmark datasets. These datasets are of different dimensions and classes. Totally eight different experiments were conducted and obtained the recognition rates in range of 90.31%–99.47% (BES database), 62.50%–78.44% (SAVEE database) and 85.83%–98.70% (SUSAS database). The obtained results convincingly prove the effectiveness of the proposed feature selection algorithm when compared to the previous works and other metaheuristic algorithms (BBO and PSO).

Physiological signal based detection of driver hypovigilance using higher order spectra

In this work, the focus is on developing a system that can detect hypovigilance, which includes both drowsiness and inattention, using Electrocardiogram (ECG) and Electromyogram (EMG) signals. Drowsiness has been manipulated by allowing the driver to drive monotonously at a limited speed for long hours and inattention was manipulated by asking the driver to respond to phone calls and short messaging services. ECG and EMG signals along with the video recording have been collected throughout the experiment. The gathered physiological signals were preprocessed to remove noise and artifacts. The hypovigilance features were extracted from the preprocessed signals using higher order spectral features. The features were classified using k Nearest Neighbor, Linear Discriminant Analysis and Quadratic Discriminant Analysis. The bispectral features gave an overall maximum accuracy of 96.75% and 92.31% for ECG and EMG signals, respectively using k fold validation. The features of ECG and EMG signals were fused using principal component analysis to obtain the optimally combined features and the classification accuracy was 96%. A number of road accidents can be avoided if an alert is sent to a driver who is drowsy or inattentive.

A physiological measures-based method for detecting inattention in drivers using machine learning approach

BackgroundIn recent years, as a result of the usage of electronic gadgets in vehicles, driver inattention has become one of the major causes of road accidents that lead to severe physical injuries, deaths and significant economic losses. Statistics ensure the need of a reliable driver inattention detection system that can alert the driver before a mishap happens. MethodsIn this work, we aimed to develop a system that can detect inattention using electrocardiogram (ECG) and surface electromyogram (sEMG) signals. Cognitive and visual inattention was manipulated by asking the driver to respond to phone calls and short messaging services, respectively. A total of 15 male subjects participated in the data collection process. The subjects were asked to drive for two hours in a simulated environment at three different times of the day. ECG, sEMG and video were obtained throughout the experiment. The gathered physiological signals were preprocessed to remove noises and artefacts. The inattention features were extracted from the preprocessed signals using conventional statistical, higher-order statistical and higher-order spectral features. The features were classified using k-nearest neighbour analysis, linear discriminant analysis and quadratic discriminant analysis. ResultsThe bispectral features gave overall maximum accuracies of 98.12% and 90.97% for the ECG and EMG signals, respectively. ConclusionWe conclude that ECG and EMG signals can be explored further to develop a robust and reliable inattention detection system.

Separating Physical Components from Galaxy Spectra by Subspace Methods

Using subspace methods, we study the distribution of physical components of galaxies in wavelength space. We find that it is valid to assume that the stellar and the gaseous components of galaxies span complementary subspaces. To first order, stellar and gaseous spectral features can be extracted from galaxy spectra through a simple matrix multiplication. By comparing the stellar continua obtained respectively using the model-based and the empirical approach through a commonality measure, we conclude that the latter may lose higher-order spectral features.

Identification of gastroenteric viruses by electron microscopy using higher order spectral features

Background: Many paediatric illnesses are caused by viral agents, for example, acute gastroenteritis. Electron microscopy can provide images of viral particles and can be used to identify the agents. Objectives: The use of electron microscopy as a diagnostic tool is limited by the need for high level of expertise in interpreting these images and the time required. A semi-automated method is proposed in this paper. Study design: The method is based on bispectal features that capture contour and texture information while providing robustness to shift, rotation, changes in size and noise. The magnification or true size of the viral particles need not be known precisely, but if available can be used additionally for improved classification. Viral particles from one or more images are segmented and analyzed to verify whether they belong to a particular class (such as Adenovirus, Rotavirus, etc.) or not. Two experiments were conducted—depending on the populations from which virus particle images were collected for training and testing, respectively. In the first, disjoint subsets from a pooled population of virus particles obtained from several images were used. In the second, separate populations from separate images were used. The performance of the method on viruses of similar size was separately evaluated using Astrovirus, HAV and Poliovirus. A Gaussian Mixture Model was used for the probability density of the features. A threshold on the log-likelihood is varied to study false alarm and false rejection trade-off. Features from many particles and/or likelihoods from independent tests are averaged to yield better performance. Results: An equal error rate (EER) of 2% is obtained for verification of Rotavirus (tested against three other viruses) when features from 15 viral particle images are averaged. It drops further to less than 0.2% when scores from two tests are averaged to make a decision. For verification of Astrovirus (tested against two others of the same size) the EER was less than 2% when 20 particles and two tests were used. Conclusion: Bispectral features and Gaussian mixture modelling of their probability density are shown to be effective in identifying viruses from electron microscope images. With the use of digital imaging in electron microscopes, this method can be fully automated.

Detection of mines in acoustic images using higher order spectral features

A new pattern-recognition algorithm detects approximately 90% of the mines hidden in the Coastal Systems Station Sonar0, 1, and 3 databases of cluttered acoustic images, with about 10% false alarms. Similar to other approaches, the algorithm presented here includes processing the images with an adaptive Wiener filter (the degree of smoothing depends on the signal strength in a local neighborhood) to remove noise without destroying the structural information in the mine shapes, followed by a two-dimensional FIR filter designed to suppress noise and clutter, while enhancing the target signature. A double peak pattern is produced as the FIR filter passes over mine highlight and shadow regions. Although the location, size, and orientation of this pattern within a region of the image can vary, features derived from higher order spectra (HOS) are invariant to translation, rotation, and scaling, while capturing the spatial correlations of mine-like objects. Classification accuracy is improved by combining features based on geometrical properties of the filter output with features based on HOS. The highest accuracy is obtained by fusing classification based on bispectral features with classification based on trispectral features.