Class Support Vector Machine Research Articles

Extracting Multiple Features for Dynamic Hand Gesture Recognition

In this work a framework based on histogram of orientation of optical flow (HOOF) and local binary pattern from three orthogonal planes (LBP_TOP) is proposed for recognizing dynamic hand gestures. HOOF algorithm extracts local shape and dynamic motion information of gestures from image sequences and local descriptor LBP is extended to three orthogonal planes to create an efficient motion descriptor. These features are invariant to scale, translation, illumination and direction of motion. The performance of the new framework is tested in two different ways. The first one is by fusing the global and local features as one descriptor and the other is using features separately to train the multi class support vector machine. Performance analysis shows that the proposed approach produces better results for recognizing dynamic hand gestures when compared with state of the art methods

Detection of European Aspen (Populus tremula L.) Based on an Unmanned Aerial Vehicle Approach in Boreal Forests

European aspen (Populus tremula L.) is a keystone species for biodiversity of boreal forests. Large-diameter aspens maintain the diversity of hundreds of species, many of which are threatened in Fennoscandia. Due to a low economic value and relatively sparse and scattered occurrence of aspen in boreal forests, there is a lack of information of the spatial and temporal distribution of aspen, which hampers efficient planning and implementation of sustainable forest management practices and conservation efforts. Our objective was to assess identification of European aspen at the individual tree level in a southern boreal forest using high-resolution photogrammetric point cloud (PPC) and multispectral (MSP) orthomosaics acquired with an unmanned aerial vehicle (UAV). The structure-from-motion approach was applied to generate RGB imagery-based PPC to be used for individual tree-crown delineation. Multispectral data were collected using two UAV cameras: Parrot Sequoia and MicaSense RedEdge-M. Tree-crown outlines were obtained from watershed segmentation of PPC data and intersected with multispectral mosaics to extract and calculate spectral metrics for individual trees. We assessed the role of spectral data features extracted from PPC and multispectral mosaics and a combination of it, using a machine learning classifier—Support Vector Machine (SVM) to perform two different classifications: discrimination of aspen from the other species combined into one class and classification of all four species (aspen, birch, pine, spruce) simultaneously. In the first scenario, the highest classification accuracy of 84% (F1-score) for aspen and overall accuracy of 90.1% was achieved using only RGB features from PPC, whereas in the second scenario, the highest classification accuracy of 86 % (F1-score) for aspen and overall accuracy of 83.3% was achieved using the combination of RGB and MSP features. The proposed method provides a new possibility for the rapid assessment of aspen occurrence to enable more efficient forest management as well as contribute to biodiversity monitoring and conservation efforts in boreal forests.

Development of MEMS accelerometer based real time data acquisition prototype for COVID-19 patient breath monitoring and assessment

In late 2019, coronavirus-2, a dangerous human pathogen that causes the disease coronavirus disease 2019, arose as a severe human pathogen (COVID-19). Acute respiratory failure, which is similar to acute respiratory distress syndrome, is the most frequent clinical presentation of extreme COVID-19. COVID-19 includes diseases of the airway, lung parenchymal, pulmonary vascular, and respiratory neuromuscular systems. This essay summarises what is learned about the impact of coronavirus-2 infection on various areas of the respiratory system, provides insights into the basic pathology of respiratory disease, and highlights current and prospective translational and clinical research. This paper discusses economic and reliable respiratory monitoring systems using prototype popular embedded controller and built-in smartphone MEMS accelerometer. The hardware is tested and implemented on healthy, and Covid-19 affected persons for respiratory data, and Correlation has been made among them. One class SVM(Support Vector Machine) learning algorithm further processes the data collected from the sensor. The project primarily aims to detect anomalies in the respiratory data set to assess the rate of respiration by matching the data sets of a healthy individual and a COVID infected person.

Video anomaly detection and localization based on appearance and motion models

In this paper, we present an approach to detect and localize anomalies in the surveillance videos. Precise detection, modeling the normality in a context and dealing with false alarms are the major challenges to cope with while performing detection and localization of anomalies. We propose appearance and motion models to detect an anomalous event at the frame level in a video. In the anomalous frames, we localize the anomaly using appearance and connected component analysis cues. The proposed appearance model utilizes the mask-rcnn for scene semantic analysis. The proposed motion model constructs frame descriptors based on the histograms of the intensity difference maps of the consecutive video frames. A one class support vector machine (OCSVM) is used to detect motion based anomalous event. In order to eliminate the false alarms of anomaly detection and localization emerging due to camera jitter and object movements in unlikely motion regions of a video frame, we exploit the connected component analysis. Experiments on the Avenue and UMN datasets reveal that the proposed approach achieves high accuracy and outperforms the existing methods.

Support vector machines with the known feature-evolution priors

In traditional SVMs (support vector machines), each feature involved in an object is assumed to contribute equally to the classification accuracy. However, the equal importance assumption for each feature is inappropriate for classifying the objects. The existing weighting strategies usually utilize the fixed scaling factors to measure the importance of each feature and hence they are not good fits for feature evolution. In this paper, with the assumption that a Known Feature-Evolution Prior (KFEP) can be employed to capture importance distribution from the perspective of both feature granularity and object granularity, SVM-based machines called KFEP-SVMs are developed to capture the importance of features dynamically. Also, it is revealed that SVMs with the known feature-evolution priors are equivalent to SVMs on a higher feature space. Besides, we theoretically demonstrate that the proposed machines KFEP-SVMs with the known feature-evolution prior indeed have enhanced generalization ability, compared with the classic SVMs. Experimental results show the effectiveness of the proposed machines KFEP-SVMs.

Microgrid protection against high impedance faults with robustness to harmonic intrusion and weather intermittency

The high impedance faults (HIF) in microgrids are characterized by low magnitude fault current with inherent non-linearity and random behavior. The possible weather intermittency in PV integrated microgrids and the intrusion of harmonics due to power converters and non-linear loading, further reduce the fault current during HIF. The insignificant fault current does not allow detection of HIF by conventional overcurrent relays. In this regard, a reliable and robust protection scheme based on the combined framework of probabilistic modelling, least square and Adaline algorithm, sine cosine algorithm (SCA) and support vector machine (SVM), has been proposed to perform the tasks of mode detection, fault detection/classification and section identification during HIF in microgrid. The algorithm involves online extraction of harmonic components and stochastic modelling of intermittency in solar irradiance. The performance of the proposed scheme has been analysed in terms of robustness against harmonics intrusion and weather intermittency during HIF. The performance comparison of the proposed scheme with the decision tree (DT) and classical SVM based techniques reflects its effectiveness with improved immunity against harmonic intrusion and weather intermittency. Further, the proposed scheme has been validated using OPAL-RT digital simulator for evaluating its appropriateness in real-time settings.

On Lagrangian L2-norm pinball twin bounded support vector machine via unconstrained convex minimization

With the introduction of the regularization term in the formulation of the well-known twin support vector machine (TWSVM) for classification , twin bounded support vector machine (TBSVM) method was proposed recently as an improved version by implementing the structural risk minimization principle. However, TBSVM employs hinge loss function and it is sensitive to noise and unstable to re-sampling. Since the pinball loss function related to quantile distance enjoys noise insensitivity property , a novel TBSVM method with squared pinball loss function for classification is proposed. The noise insensitivity and scatter minimization properties are discussed. Our formulation is further simplified as a pair of unconstrained strongly convex minimization problems in the dual space free of matrix inversion terms and having only m variables where m is the number of training examples. As opposed to TWSVM and TBSVM wherein approximate kernel generated surfaces are constructed, kernel trick is applied directly in our formulation and thereby elegant formulation as in the classical support vector machine (SVM) is achieved. Numerical experiments performed on a synthetic and thirteen benchmark datasets with noise where better or comparable generalization performance with faster learning speed by the proposed method confirms its suitability and applicability to problems of interest.

Gradient Boosting Machine Based on PSO for prediction of Leukemia after a Breast Cancer Diagnosis

The purpose of this study is to develop an accurate risk predictive model for Chronic Myeloid Leukemia (CML) after an early diagnosis of Breast Cancer (BC). Gradient Boosting Machine (GBM) classification algorithm has been applied to the SEER breast cancer dataset for females diagnosed with BC from 2010 to 2016. A practical Swarm optimizer (PSO) was utilized to optimize the GBM algorithm's hyperparameters to find the SEER dataset's best attributes. Nine attributes were carefully selected to study the growth of CML after a lag time of 6 months following BC's diagnosis. The results revealed that the predictive model could classify patients with breast cancer only and patients with breast cancer with Leukemia by an achieved Accuracy, Sensitivity, and Specificity rates of 98.5 %, 99 %, 97.85 %, respectively. To verify the performance of the proposed algorithm, the accuracy of the suggested GBM classifier model was compared with another state-of-the-art model classifiers KNN (k-Nearest Neighbor), SVM (Support Vector Machine), and RF (Random Forest), which are commonly applied algorithms in most of the existing literature. The results also proved the superior ability of the implemented GBM model Classifier in the classification of breast cancer disease and prediction of patients having Leukemia developed after having breast cancer. These results are promising as they show the integral role of the GBM classifier to classify and predict the tumor with high accuracy and efficiency, which will further help in better cancer diagnosis and treatment of the disease.

Deep Learning Based Sea Ice Classification with Gaofen-3 Fully Polarimetric SAR Data

In this paper, the performance of C-band synthetic aperture radar (SAR) Gaofen-3 (GF-3) quad-polarization Stripmap (QPS) data is assessed for classifying late spring and summer sea ice types. The investigation is based on 18 scenes of GF-3 QPS data acquired in the Arctic Ocean in 2017. In this study, floe ice (FI), brash ice (BI) between floes and open water (OW, ice-free area) were classified based on a mini sea ice residual convolutional network, which we call MSI-ResNet. While investigating the optimal patch size for MSI-ResNet, we found that, as the patch size continues to grow, the classification accuracy first increases and then decreases. A patch size of 31 × 31 was found to achieve the best performance. The performance of classification using different polarization combinations from the QPS data was also assessed. The vertical-vertical (VV) polarization input overestimates the FI category while incorrectly identifying most of the BI as FI. The VH polarization produces a synchronous improvement in FI, BI, and OW discrimination, with a higher overall accuracy and kappa coefficient (91.09% and 0.85, respectively) than the VV polarization (83.37% and 0.70, respectively). The combination of VV and vertical-horizontal (VH) polarizations presents a modest precision improvement for BI and OW together with a slight overestimation for FI. With VV, VH, and horizontal-horizontal (HH) polarization data as the inputs, the user’s accuracy improves to 95.12%, 93.42%, and 95.17% for FI, BI, and OW, respectively. The accuracy was assessed against visual interpretation of the sea ice classes in the images using a stratified sampling method. The application of the MSI-ResNet method to data covering the Beaufort Sea and the north of the Severnaya Zemlya archipelago was found to achieve a high overall accuracy (kappa) of 94.62% (±0.92) and 94.23% (±0.90), respectively. This is similar to the classification accuracy obtained in the Fram Strait. From the results of this study, it is shown that the MSI-ResNet method performs better than the classical support vector machine (SVM) classifier for sea ice discrimination. The GF-3 QPS mode data also show more details in discriminating scattered sea ice floes than the coincident Sentinel-1A Extra Wide (EW) swath mode data.

A Novel Secure Data Hiding Technique into Video Sequences Using RVIHS

Most of the present hiding techniques on video are considered over plaintext domain and plain video sequences are used to embed information bits. The work presented here reveals the novelty for information embedding in a video sequence over the ciphered domain. The carrier video signal is encrypted using chaos technique which uses multiple chaotic maps for encryption. The proposed reversible video information hiding scheme (RVIHS) exhibits an innovative property that, at the decoding side we can perfectly extract the information along with carrier video without any distortion. The public key modulation is a mechanism used to achieve data embedding, where as in secret key encryption is not required. The proposed approach is used to differentiate encoded and non-encoded picture patches at decoder end by implementing 2 class Support Vector Machine grouping. This helps for us to retrieve the original visual sequence with embedded message and to scale up embedding capacity. The experiment is conducted using real time videos for embedding the information. The outcome of proposed work bring about best embedding capacity, compared to existing techniques.

Artificial Intelligence for Damage Detection in Automotive Composite Parts: A Use Case

<div class="section abstract"><div class="htmlview paragraph">The detection and evaluation of damage in composite materials components is one of the main concerns for automotive engineers. It is acknowledged that defects appeared in the manufacturing stage or due to the impact and/or fatigue loads can develop along the vehicle riding. To avoid an unexpected failure of structural components, engineers ask for cheap methodologies assessing the health state of composite parts by means of continuous monitoring. Non Destructive Technique (NDT) for the damage assessment of composite structures are nowadays common and accurate, but an on-line monitoring requires properties as low cost, small size and low power that do not belong to common NDT. The presence of a damage in composite materials, either due to fatigue cycling or low-energy impact, leads to progressive degradation of elastic moduli and strengths. Since there is a well-known relationship between the elastic modulus reduction and the amount of damage, the stiffness degradation can be used for the scope of detecting the position and the amount of damage that has taken place. Relying on these concepts, a novel strain-based damage sensing procedure is here proposed, that can identify damages in composite structures by processing strain measures from a distributed sensors network. To achieve this result a combined Machine Learning pipeline, composed by Principal Component Analysis (PCA) and One Class Support Vector Machine (OC-SVM) is proposed. First, PCA learns a linear transformation on the undamaged measurements to reduce the data dimensionality; secondly, OC-SVM trained to detect anomalies in the projected components. A cross-validation procedure is used to find the optimal pipeline configuration. The methodology is virtually tested on a carbon fiber suspension. The results suggest dropping the first components of the PCA to feed the classifier. In addition, results show the capability of the algorithm to detect anomalies in the component strain response.</div></div>

Discrete cosine transform and multi class support vector machines for classification cardiac atrial arrhythmia and cardiac normal

The electrocardiogram signal is the most important analysis to detect cardiac arrhythmia. Machine learning classification is used as a first step to detect someone’s arrhythmia or normal heart. This paper discusses one method for detecting arrhythmia by using digital images of cardiac signals and R-R intervals. The process electrocardiogram digital image is divided into two, first the process of calculating the R-R intervals and second the process of extraction feature using Discrete Cosine Transform, followed by calculating the Euclidean Distance or Cityblock Distance with normal electrocardiogram signal reference. Euclidean Distance results or Cityblock Distance and R-R distance of electrocardiogram signals are then classified using Multiclass Support Vector Machine. The results of accuracy the classification four classes that are cardiac normal, atrial premature beat arrhythmia, atrial flutter arrhythmia, and atrial fibrillation arrhythmia, are 81.9%. The originality is used image to detect cardiac normal or cardiac arrhythmia by combined Discrete Cosine Transform, Euclidean distance or City block distance and Multiclass Support Vector Machine.

A hybrid deep network based approach for crowd anomaly detection

In this paper, we present a hybrid deep network based approach for crowd anomaly detection in videos. For improved performance, the proposed approach exploits deep and handcrafted features. The proposed approach extracts spatial and temporal deep features from video frames using two resnet101 models. In order to enhance the deep features discrimination between normal and anomalous events, we perform smoothing of their Euclidean distance values for consecutive frames. For a handcrafted feature that describes the high level motion at the frame level, we compute gradient sum of the frame difference of consecutive video frames. Two deep features and one handcrafted feature of the training frames are used to train three one class support vector machines (OCSVMs). A frame is classified as anomalous performing decision combination of three OCSVMs. Experiments reveal that the proposed approach achieves high accuracy on UMN crowd anomaly dataset. On a more challenging PETS 2009 dataset the proposed approach achieves comparable performance to existing approaches.

One-class graph neural networks for anomaly detection in attributed networks

Nowadays, graph-structured data are increasingly used to model complex systems. Meanwhile, detecting anomalies from graph has become a vital research problem of pressing societal concerns. Anomaly detection is an unsupervised learning task of identifying rare data that differ from the majority. As one of the dominant anomaly detection algorithms, One Class Support Vector Machine has been widely used to detect outliers. However, those traditional anomaly detection methods lost their effectiveness in graph data. Since traditional anomaly detection methods are stable, robust and easy to use, it is vitally important to generalize them to graph data. In this work, we propose One Class Graph Neural Network (OCGNN), a one-class classification framework for graph anomaly detection. OCGNN is designed to combine the powerful representation ability of Graph Neural Networks along with the classical one-class objective. Compared with other baselines, OCGNN achieves significant improvements in extensive experiments.

A method of non-bug report identification from bug report repository

One of the most common issues addressed by bug report studies is misclassification when identifying and then filtering non-bug reports from the bug report repository. Having to filter out unrelated reports wastes time in identifying actual bug reports, and this escalates costs as extra maintenance and effort are required to triage and fix bugs. Therefore, this issue has been seriously studied and is addressed here. To tackle this problem, this study proposes a method of automatically identifying non-bug reports in the bug report repository using classification techniques. Three points are considered here. First, the bug report features used are unigram and CamelCase, where CamelCase words are used for feature expansion. Second, five term weighting schemes are compared to determine an appropriate term weighting scheme for this task. Lastly, the support vector machine (SVM) family i.e. binary-class SVM, one class SVM based on Scholkopf methodology and support vector data description (SVDD) are used as the main mechanisms for modeling non-bug report identifiers. After testing by recall, precision, and F1, the results demonstrate the efficiency of identifying non-bug reports in the bug report repository. Our results may be acceptable after comparing to the previous well-known studies, and the performance of non-bug report identifiers with tf-igm and modified tf-icf weighting schemes for both Scolkopf methodology and SVDD methods yielded the best value when compared to others.

Motor Training Using Mental Workload (MWL) With an Assistive Soft Exoskeleton System: A Functional Near-Infrared Spectroscopy (fNIRS) Study for Brain-Machine Interface (BMI).

Mental workload is a neuroergonomic human factor, which is widely used in planning a system's safety and areas like brain–machine interface (BMI), neurofeedback, and assistive technologies. Robotic prosthetics methodologies are employed for assisting hemiplegic patients in performing routine activities. Assistive technologies' design and operation are required to have an easy interface with the brain with fewer protocols, in an attempt to optimize mobility and autonomy. The possible answer to these design questions may lie in neuroergonomics coupled with BMI systems. In this study, two human factors are addressed: designing a lightweight wearable robotic exoskeleton hand that is used to assist the potential stroke patients with an integrated portable brain interface using mental workload (MWL) signals acquired with portable functional near-infrared spectroscopy (fNIRS) system. The system may generate command signals for operating a wearable robotic exoskeleton hand using two-state MWL signals. The fNIRS system is used to record optical signals in the form of change in concentration of oxy and deoxygenated hemoglobin (HbO and HbR) from the pre-frontal cortex (PFC) region of the brain. Fifteen participants participated in this study and were given hand-grasping tasks. Two-state MWL signals acquired from the PFC region of the participant's brain are segregated using machine learning classifier—support vector machines (SVM) to utilize in operating a robotic exoskeleton hand. The maximum classification accuracy is 91.31%, using a combination of mean-slope features with an average information transfer rate (ITR) of 1.43. These results show the feasibility of a two-state MWL (fNIRS-based) robotic exoskeleton hand (BMI system) for hemiplegic patients assisting in the physical grasping tasks.

A Hybrid Supervised Machine Learning Classifier System for Breast Cancer Prognosis Using Feature Selection and Data Imbalance Handling Approaches

Nowadays, breast cancer is the most frequent cancer among women. Early detection is a critical issue that can be effectively achieved by machine learning (ML) techniques. Thus in this article, the methods to improve the accuracy of ML classification models for the prognosis of breast cancer are investigated. Wrapper-based feature selection approach along with nature-inspired algorithms such as Particle Swarm Optimization, Genetic Search, and Greedy Stepwise has been used to identify the important features. On these selected features popular machine learning classifiers Support Vector Machine, J48 (C4.5 Decision Tree Algorithm), Multilayer-Perceptron (a feed-forward ANN) were used in the system. The methodology of the proposed system is structured into five stages which include (1) Data Pre-processing; (2) Data imbalance handling; (3) Feature Selection; (4) Machine Learning Classifiers; (5) classifier’s performance evaluation. The dataset under this research experimentation is referred from the UCI Machine Learning Repository, named Breast Cancer Wisconsin (Diagnostic) Data Set. This article indicated that the J48 decision tree classifier is the appropriate machine learning-based classifier for optimum breast cancer prognosis. Support Vector Machine with Particle Swarm Optimization algorithm for feature selection achieves the accuracy of 98.24%, MCC = 0.961, Sensitivity = 99.11%, Specificity = 96.54%, and Kappa statistics of 0.9606. It is also observed that the J48 Decision Tree classifier with the Genetic Search algorithm for feature selection achieves the accuracy of 98.83%, MCC = 0.974, Sensitivity = 98.95%, Specificity = 98.58%, and Kappa statistics of 0.9735. Furthermore, Multilayer Perceptron ANN classifier with Genetic Search algorithm for feature selection achieves the accuracy of 98.59%, MCC = 0.968, Sensitivity = 98.6%, Specificity = 98.57%, and Kappa statistics of 0.9682.

Local Influence in Constrained General Linear Models

support vector machines (SVMs) constitute one of the most popular and powerful classification methods. However, SVMs can be limited in their performance on highly imbalanced datasets. A classifier which has been trained on an imbalanced dataset can produce a biased model towards the majority class and result in high misclassification rate for minority class. For many applications, especially for medical diagnosis, it is of high importance to accurately distinguish false negative from false positive results. The purpose of this study is to successfully evaluate the performance of a classifier, keeping the correct balance between sensitivity and specificity, in order to enable the success of trauma outcome prediction. We compare the standard (or classic) SVM (C SVM) with resampling methods and a cost sensitive method, called Two Cost SVM (TC SVM), which constitute widely accepted strategies for imbalanced datasets and the derived results were discussed in terms of the sensitivity analysis and receiver operating characteristic (ROC) curves.

Multi class SVM algorithm with active learning for network traffic classification

With the current massive amount of traffic that is going through the internet, internet service providers (ISPs) and networking service providers (NSPs) are looking for various ways to accurately predict the application type of flow that is going through the internet. Such prediction is critical for security and network monitoring applications as they require application type to be known in prior. Traditional ways using port-based or payload-based analysis are not sufficient anymore as many applications start using dynamic unknown port numbers, masquerading, and encryption techniques to avoid being detected. Recently, machine learning has gained significant attention in many prediction applications including traffic classification from flow features or characteristics. However, such algorithms suffer from an imbalanced data problem where some applications have fewer flow data and hence difficult to predict. In this paper, we employ network flow-level characteristics to identify the application type of traffic. Furthermore, we propose the use of an improved support vector machine (SVM) algorithm, named cost-sensitive SVM (CMSVM), to solve the imbalance problem in network traffic identification. CMSVM adopts a multi-class SVM algorithm with active learning which dynamically assigns a weight for applications. We examine the classification accuracy and performance of the CMSVM algorithm using two different datasets, namely MOORE_SET and NOC_SET datasets. Our results show that the CMSVM algorithm can reduce computation cost, improve classification accuracy and solve the imbalance problem when compared to other machine learning techniques.

Evaluating satellite hyperspectral (Orbita) and multispectral (Landsat 8 and Sentinel-2) imagery for identifying cotton acreage

ABSTRACT Hyperspectral remote sensing demonstrates a great potential for identifying crops. Orbita hyperspectral image satellite (OHS) is a new hyperspectral satellite in orbit with the highest spectral and spatial resolutions and relatively short spectral range. In this study, the OHS, along with Land Remote-Sensing Satellite (System, Landsat) eight operational land imager (OLI) images (assigned as L8) and Sentinel-2B (S2B) images, was used to identify cotton and estimate its acreage in Beiquan town, Shihezi city, Xinjiang Uygur Autonomous Region of China based on spectral reflectance, sample separability, and accuracy. And Two Class Support vector machine was adapted to map cotton acreage. For the spectral reflectance of cotton, OHS is slightly higher than the analytical spectral device (ASD) in the visible range and is similar in the near-infrared range. Both the S2B and L8 spectrum curves of cotton fit well with the ASD spectrum’s. For the sample separability, the pair values between cotton and non-cotton (e.g. impervious surface, water, maize, grape, etc.) of OHS and S2B are better than that of L8 in general, indicating that OHS and S2B are more reliable than L8 for cotton identification using OHS, L8, and S2B. The results show that the overall accuracy of these three sensors is comparable: an overall accuracy of 0.96 from OHS, 0.95 from S2B, and 0.94 from L8. Moreover, the cotton identification accuracy is not only related to the spatial resolution but also to the spectral resolution and band position, and more importantly to each sensor specifications, even though each satellite hyperspectral and multispectral sensor has its own unique advantages in being complimentary for identifying cotton and other crops.