Multi-class Support Vector Machine Research Articles

Experimental Investigation of Body-Centric Indoor Localization Using Compact Wearable Antennas and Machine Learning Algorithms

A simple and effective body-centric localization algorithm has been proposed in this article using ultra-wideband (UWB) wearable technology. The algorithm has been validated through measurement campaigns with a human subject volunteer in an indoor environment. The algorithm takes into account statistical channel parameter analysis, machine learning (ML) algorithms, and time-of-arrival (TOA)-based range estimation/data fusion techniques. Two channel parameters namely path loss magnitude and rms delay spread are proposed as classification features to be applied to the ML algorithm to accurately classify the off-body channel links into LOS, PNLOS, and NLOS scenarios. Multiclass support vector machine (MC-SVM) classifier along with SMOTE algorithm to take into account class imbalance is applied with the channel classification accuracy of 98.63%. Threshold-based range estimation algorithms are applied in order to mitigate NLOS scenarios caused mainly due to the presence of the human subject. Results report human localization accuracy in 0.5–3 cm range using TDOA data fusion technique for target estimation. Further validation is presented considering wide range of Tx–Rx distance, presence of another obstruction between the Tx and Rx links, and performance in different environment which shows the suitability of the proposed methodology.

Experimental Validation of Speech Improvement-Based Stratified Adaptive Finite-Time Saturation Control of Omnidirectional Service Robot

To implement the human-robot interactions in a noisy environment, the speech improvement-based (SIB) stratified adaptive finite-time saturation control (SAFTSC) for omnidirectional service robot (OSR) is developed. From the outset, the feature vectors of nine designed speech commands are extracted from their frequency signals and then trained by multiclass support vector machine. Two background noises are on-line filtered with the characteristic: ``the smaller error in power spectrum is, the larger recovery from noisy power spectrum is.'' Comparisons among without or with noise, and filtering are addressed. To achieve the zero pose error of OSR in finite time, an adaptive finite-time indirect trajectory (AFTIT) is constructed. To track the AFTIT with the zero error in finite time, the adaptive finite-time saturation control (AFTSC) is also established. Both AFTIT and AFTSC possess nonlinear switching gain increasing the high-frequency motion capability to fulfill the classified speech command. Simply put, the proposed SIB stratified AFTSC includes the speech improvement for classification, the AFTIT, and the AFTSC. Besides the stability of the closed-loop system is verified by the Lyapunov stability theory, three categories of SIB experiments are compared.

Analysis of E-tongue data for tea classification based on semi-supervised learning of generative adversarial network

The identification of tea quality provides the protection of rights and interests to consumers. This paper proposes a semi-supervised learning model based on a generative adversarial network (GAN-SSL), to model and predict the electronic tongue data of tea samples. The GAN-SSL consists of generative model G and discriminative model D . We demonstrated the generated sample as a new class. The model G was inputted into a random noise. The discriminator acquires a series of potential features of the inputted data. Through the adversarial learning between the D and G , more realistic samples were generated, and D predicted the category and authenticity of the real data and generated data. The analysis of collected E-tongue data set of tea showed that GAN-SSL effectively improved the accuracy of classification compared to multi-class support vector machine (multi-class SVM), partial least square regression analysis (PLS-DA) and Decision Tree.

Multi-Class Brain Disease Classification Using Modified Pre-Trained Convolutional Neural Networks Model with Substantial Data Augmentation

The integration of various algorithms in the medical field to diagnose brain disorders is significant. Generally, Computed Tomography, Magnetic Resonance Imaging techniques have been used to diagnose brain images. Subsequently, segmentation and classification of brain disease remain an exigent task in medical image processing. This paper presents an extended model for brain image classification based on a Modified pre-trained convolutional neural network model with extensive data augmentation. The proposed system has been efficiently trained using the technique of substantial data augmentation in the pre-processing stage. In the first phase, the pre-trained models namely AlexNet, VGGNet-19, and ResNet-50 are employed to classify the brain disease. In the second phase, the idea of integrating the existing pre-trained model with a multiclass linear support vector machine is incorporated. Hence, the SoftMax layer of pre-trained models is replaced with a multi class linear support vector machine classifier is proposed. These proposed modified pre-trained model is employed to classify brain images as normal, inflammatory, degenerative, neoplastic and cerebrovascular diseases. The training loss, mean square error, and classification accuracy have been improved through the concept of Cyclic Learning rate. The appropriateness of transfer learning has been demonstrated by applying three convolutional neural network models, namely, AlexNet, VGGNet-19, and ResNet-50. It has been observed that the modified pre-trained models achieved a higher classification rate of accuracies of 93.45% when compared with a finetuned pre-trained model of 89.65%. The best classification accuracy of 92.11%, 92.83% and 93.45% has been attained in the proposed method of the modified pre-trained model. A comparison of the proposed model with other pre-trained models is also presented.

Multi-class SVM based C3D Framework for Real-Time Anomaly Detection

The conventional multi-class anomaly detection models are independent of noise elimination and feature segmentation due to large number of feature space and training images. As the number of human anomaly classes is increasing, it is difficult to find the multi-class anomaly due to high computational memory and time. In order to improve the multi-class human anomaly detection process, an advanced multi-class segmentation-based classification model is designed and implemented on the different human anomaly action databases. In the proposed model, a hybrid filtered based C3D framework is used to find the essential key features from the multiple human action data and an ensemble multi-class classification model is implemented in order to predict the new type of actions with high accuracy. Experimental outcomes proved that the proposed multi- class classification C3D model has better human anomaly detection rate than the traditional multi-class segmentation models.

Sentiment Analysis using Optimized Feature Sets in Different Facebook/Twitter Dataset Domains using Big Data

Now days, in many real life applications, the sentiment analysis plays very vital role for automatic prediction of human being activities especially on online social networks (OSNs). Therefore since from last decade, the research on opinion mining and sentiment analysis is growing with increasing volume of online reviews available over the social media networks like Facebook OSNs. Sentiment analysis falls under the data mining domain research problem. Sentiment analysis is kind of text mining process used to determine the subjective attitude like sentiment from the written texts and hence becoming the main research interest in domain of natural language processing and data mining. The main task in sentiment analysis is classifying human sentiment with objective of classifying the sentiment or emotion of end users for their specific text on OSNs. There are number of research methods designed already for sentiment analysis. There are many factors like accuracy, efficiency, speed etc. used to evaluate the effectiveness of sentiment analysis methods. The MapReduce framework under the domain of big-data is used to minimize the speed of execution and efficiency recently with many data mining methods. The sentiment analysis for Facebook OSNs messages is very challenging tasks as compared to other sentiment analysis because of misspellings and slang words presence in twitter dataset. In this paper, different solutions recently presented are discussed in detail. Then proposed the new approach for sentiment analysis based on hybrid features extraction methods and multi-class Support Vector Machine (SVM). These algorithms are designed using the Big-data techniques to optimize the performance of sentiment analysis

Integration of gene co-expression analysis and multi-class SVM specifies the functional players involved in determining the fate of HTLV-1 infection toward the development of cancer (ATLL) or neurological disorder (HAM/TSP).

Human T-cell Leukemia Virus type-1 (HTLV-1) is an oncovirus that may cause two main life-threatening diseases including a cancer type named Adult T-cell Leukemia/Lymphoma (ATLL) and a neurological and immune disturbance known as HTLV-1 Associated Myelopathy/Tropical Spastic Paraparesis (HAM/TSP). However, a large number of the infected subjects remain as asymptomatic carriers (ACs). There is no comprehensive study that determines which dysregulated genes differentiate the pathogenesis routes toward ATLL or HAM/TSP. Therefore, two main algorithms including weighted gene co-expression analysis (WGCNA) and multi-class support vector machines (SVM) were utilized to find major gene players in each condition. WGCNA was used to find the highly co-regulated genes and multi-class SVM was employed to identify the most important classifier genes. The identified modules from WGCNA were validated in the external datasets. Furthermore, to find specific modules for ATLL and HAM/TSP, the non-preserved modules in another condition were found. In the next step, a model was constructed by multi-class SVM. The results revealed 467, 3249, and 716 classifiers for ACs, ATLL, and HAM/TSP, respectively. Eventually, the common genes between the WGCNA results and classifier genes resulted from multi-class SVM that also determined as differentially expressed genes, were identified. Through these step-wise analyses, PAIP1, BCAS2, COPS2, CTNNB1, FASLG, GTPBP1, HNRNPA1, RBBP6, TOP1, SLC9A1, JMY, PABPC3, and PBX1 were found as the possible critical genes involved in the progression of ATLL. Moreover, FBXO9, ZNF526, ERCC8, WDR5, and XRCC3 were identified as the conceivable major involved genes in the development of HAM/TSP. These genes can be proposed as specific biomarker candidates and therapeutic targets for each disease.

Robust GMM least square twin K-class support vector machine for urban water pipe leak recognition

Monitoring pipe operation status is very important in saving water resources and realizing sustainability of the water supply system. Currently, the support vector machine (SVM) and its improved algorithms have been used to detect leaks in pipe networks. Among them, the least square twin multi-class support vector machine (LST-KSVC) is a novel multi-classification method. However, LST-KSVC assigns same weights to leak samples, including outliers that affect it. In addition, through the quadratic loss function, LST-KSVC often misclassifies rest class. Therefore, to overcome these two drawbacks, we propose a weighted version of LST-KSVC, referred to as the GMM least square twin K-class support vector machine (GLT-KSVC). Based on the Gaussian mixture model (GMM) method, GLT-KSVC assigns larger weights to main leak samples, and assigns smaller weights to outliers, making it insensitive to outliers. Moreover, GLT-KSVC avoids the misclassification drawback for rest samples by using weighted least squares linear loss function. The leak recognition experiment revealed that GLT-KSVC is better than LST-KSVC in classification accuracy, and its calculation time is only slightly higher than that of LST-KSVC.

Real-time mental stress detection technique using neural networks towards a wearable health monitor

In recent times, detection of mental stress using physiological signals has received widespread attention from technology researchers. Although many interesting research works have been reported in this area, evidence of hardware implementation is occasional. The main challenge in stress detection research is use of the optimum number of physiological signals and real-time detection with low-complexity algorithms. In this work, a real-time stress detection technique is presented which utilizes only a photoplethysmogram (PPG) signal to achieve improved accuracy over multi-signal-based mental stress detection techniques. A short 5 s segment of a PPG signal was used for feature extraction with an autoencoder (AE), and features were minimized using recursive feature elimination (RFE) integrated with a multi-class support vector machine (SVM) classifier. The proposed AE-RFE-SVM-based mental stress detection technique was tested with the WeSAD dataset to detect four levels of mental state, namely baseline, amusement, meditation and stress, and achieved an overall accuracy, F1 score and sensitivity of 99%, 0.99 and 98%, respectively, for 5 s PPG data. The technique provided improved performance over discrete wavelet transformation-based feature extraction followed by classification with either of the five types of classifiers, namely SVM, random forest, k-nearest neighbor, linear regression and decision tree. The technique was translated into quad-core-based standalone hardware (1.2 GHz, 1 GB RAM). The resultant hardware prototype achieved low latency (∼0.4 s) and had a low memory requirement (∼1.7 MB). The present technique can be extended to develop a remote healthcare system using wearable sensors.

Machine Learning Techniques for Human Age and Gender Identification Based on Teeth X-Ray Images.

The use of digital medical images is increasing with advanced computational power that has immensely contributed to developing more sophisticated machine learning techniques. Determination of age and gender of individuals was manually performed by forensic experts by their professional skills, which may take a few days to generate results. A fully automated system was developed that identifies the gender of humans and age based on digital images of teeth. Since teeth are a strong and unique part of the human body that exhibits least subject to risk in natural structure and remains unchanged for a longer duration, the process of identification of gender- and age-related information from human beings is systematically carried out by analyzing OPG (orthopantomogram) images. A total of 1142 digital X-ray images of teeth were obtained from dental colleges from the population of the middle-east part of Karnataka state in India. 80% of the digital images were considered for training purposes, and the remaining 20% of teeth images were for the testing cases. The proposed gender and age determination system finds its application widely in the forensic field to predict results quickly and accurately. The prediction system was carried out using Multiclass SVM (MSVM) classifier algorithm for age estimation and LIBSVM classifier for gender prediction, and 96% of accuracy was achieved from the system.

Intelligent Fault Diagnosis Framework for Modular Multilevel Converters in HVDC Transmission.

Open circuit failure mode in insulated-gate bipolar transistors (IGBT) is one of the most common faults in modular multilevel converters (MMCs). Several techniques for MMC fault diagnosis based on threshold parameters have been proposed, but very few studies have considered artificial intelligence (AI) techniques. Using thresholds has the difficulty of selecting suitable threshold values for different operating conditions. In addition, very little attention has been paid to the importance of developing fast and accurate techniques for the real-life application of open-circuit failures of IGBT fault diagnosis. To achieve high classification accuracy and reduced computation time, a fault diagnosis framework with a combination of the AC-side three-phase current, and the upper and lower bridges’ currents of the MMCs to automatically classify health conditions of MMCs is proposed. In this framework, the principal component analysis (PCA) is used for feature extraction. Then, two classification algorithms—multiclass support vector machine (SVM) based on error-correcting output codes (ECOC) and multinomial logistic regression (MLR)—are used for classification. The effectiveness of the proposed framework is validated by a two-terminal simulation model of the MMC-high-voltage direct current (HVDC) transmission power system using PSCAD/EMTDC software. The simulation results demonstrate that the proposed framework is highly effective in diagnosing the health conditions of MMCs compared to recently published results.

Computer-assisted grading of follicular lymphoma: a classification based on SVM, machine learning, and transfer learning approaches

ABSTRACT This proposed work implements various classification approaches to classify the H&E-stained Follicular Lymphoma tissue sample. As part of the process, k-means clustering is used to isolate cytological components like the nucleus extracellular and cytoplasmic regions from images. Then feature space vector is constructed by combining global feature extraction techniques like LBP, LDP, GLCM, and other local feature-extraction techniques. To classify FL images into their respective grades, feature space vectors obtained from feature extraction algorithms are input into a multiclass SVM. Moreover, classification accuracies were explicitly tested with different classifiers like CNN and other pre-trained deep learning networks that can directly operate on raw images without any preprocessing to classify the FL images into their respective grades. The efficacy of different classifiers is presented. This preliminary work provides a proof of concept for incorporating automated FL tissue diagnostic systems into future pathology workflows to supplement the pathologists'.

Machine Learning Assisted Methodology for Multiclass Classification of Malignant Brain Tumors

Analysis of malignant and non-malignant brain tumors is done using a computer-aided diagnosis system by practitioners worldwide. Radiologists refer computer-assisted techniques to draw conclusions using image modalities and inferences. Pedagogically, various machine learning approaches have been used, which usually focus on the classification of imaging modality into two categories, either normal and abnormal images or differentiating between benign and malignant tumors. Still, the work requirement is to classify these multi-class malignant tumors into their specific class with better precision. The proposed work focuses on distinguishing between the types of high-grade malignant brain tumors. This study is performed on real-life malignant brain tumor datasets having five classes. The proposed methodology uses the vast feature set from six domains to capture most of the hidden information in the extracted region of interest. Later, relevant features are extracted from the feature set pool using a new proposed feature selection algorithm named the Cumulative Variance method (CVM). Next, the selected features are used for model training and testing using K-Nearest Neighbour (KNN), multi-class Support Vector Machine (mSVM), and Neural Network (NN) for predicting multi-class classification accuracy. The experiments are performed using the proposed feature selection algorithm with three classifiers. The mean average classification accuracy achieved by using the proposed approach is88.43% (KNN), 92.5% (mSVM), and 95.86% (NN), respectively. The comparative analysis of the proposed approach with other existing algorithms like ICA, and GA suggest that the proposed approach gains an increase of accuracy around 2% (KNN), 3% (SVM), and 4% (NN).The experimentation results concluded that the proposed approach is found better with NN classifier with an accuracy of 95.86% using diversified features.

Recognizing Continuous Multiple Degrees of Freedom Foot Movements With Inertial Sensors.

Recognition of continuous foot motions is important in robot-assisted lower limb rehabilitation, especially in prosthesis and exoskeleton design. For instance, perceiving foot motion is essential feedback for the robot controller. However, few studies have focused on perceiving multiple-degree of freedom (DOF) foot movements. This paper proposes a novel human-machine interaction (HMI) recognition wearable system for continuous multiple-DOF ankle-foot movements. The proposed system uses solely kinematic signals from inertial measurement units and multiclass support vector machines by creating error-correcting output codes. We conducted a study with multiple participants to validate the performance of the system using two strategies, a general model and a subject-specific model. The experimental results demonstrated satisfactory performance. The subject-specific approach achieved 98.45% ± 1.17% (mean ± SD) overall accuracy within a prediction time of 10.9 ms ± 1.7 ms, and the general approach achieved 85.3% ± 7.89% overall accuracy within a prediction time of 14.1 ms ± 4.5 ms. The results prove that the proposed system can more effectively recognize multiple continuous DOF foot movements than existing strategies. It can be applied to ankle-foot rehabilitation and fills the HMI high-level control demand for multiple-DOF wearable lower-limb robotics.

A Novel Approach for Detection of Pavement Crack and Sealed Crack Using Image Processing and Salp Swarm Algorithm Optimized Machine Learning

During the phase of periodic survey, sealed crack and crack in asphalt pavement surface should be detected accurately. Moreover, the capability of identifying these two defects can help reduce the false‐positive rate for pavement crack detection. Because crack and sealed crack are both line‐based defects and may resemble each other in shape, this study puts forward an innovative method based on computer vision for detecting sealed crack and crack. This method is an integration of feature extraction based on image processing and metaheuristic optimized machine learning. Image processing is used to compute features that characterize visual appearance and texture of the pavement image. Subsequently, Salp Swarm Algorithm integrated with multiclass support vector machine is employed for pattern recognition. Based on experimental results, the newly developed method has achieved the most desired predictive performance with an accuracy rate = 91.33% for crack detection and 92.83% for sealed crack detection.

Source Printer Identification with Microscopic Printing using Deep Learning

Due to the advent of digital technologies, it has become much easier to falsify printed documents for malicious purposes. Therefore, it is essential to research and develop efficient algorithms to distinguish authentic printed documents from fakes. A source printer identification technique is one of the methods to trace the source of the documents, thereby protecting the reliability and integrity of printed documents. This paper proposes a method to identify the source printers using Deep Learning approach. Particularly, we utilise a Convolutional Neural Network for features extraction from printed microscopic patterns of different printing sources. These extracted features are fed into a multi-class Support Vector Machine and Random Forest classification for the source identification. The actual printed dot patterns from common printing technologies such as conventional offset, waterless offset, electrophotographic are used to evaluate the effectiveness of our proposed approach. This study highlights the fact that printed patterns in gray images and features based on Convolutional Neural Network extraction help to reserve much more distinguishable characteristics of printing sources. The hybrid classification algorithms outperform the previous studies using hand-extracted features, with more than 99 % of samples correctly classified in the test set.

Classification of Pneumonia Using Deep Convolutional Neural Network

Pneumonia is considered a serious and fatal disease worldwide. In fact, pneumonia can be an individual's life-endangering if not treated promptly by drugs. Therefore, the early detection of pneumonia enhances the chances of recovery, which helps reduce mortality. It is worth noting that X-rays are one of the most important diagnostic tools for diagnosing pneumonia. In fact, Chest X-ray is widely used in the diagnosis of many lung diseases (such as: Breast Cancer, Pneumonia, Tuberculosis, etc.), due to lower diagnostic costs. Indeed, the diagnoses can be subjective for many reasons for example the appearance of disease which can be unclear in chest X-ray images or can be confused with other diseases. Hence, for enhancing the level of diagnosis to guide clinicians, computer-aided diagnosis systems will be needed. In this paper, we put forward to develop a structure to classify pneumonia from chest X-ray images using a Convolutional Neural Network (CNN) and residual network architecture. Clearly, to determine if a person is infected with pneumonia or not, we used two well-known CNN pre-trained models (ResNet50 and ResNet101), with multi-class Support Vector Machine (SVM) to classify and transfer learning from the pre-trained CNN models to extract and classify features. Thus, the proposed framework takes an X-ray image size of 224 x 224 pixels as an input and gives the final prediction Normal or Pneumonia. The experimental results showed that the classification models proved to be effective, with an accuracy range of 97% to 98.3%. More precisely, the image extraction features using Resnet50 + SVM and Transfer Learning + Resnet50 methods achieve the highest performance of Accuracy of 98.3% and 97.8%, respectively.

Programmed for Automatic Bone Disorder Clustering Based on Cumulative Calcium Prediction for Feature Extraction.

The prediction of bone disorders varies between ortho-physicians. A precise bone disorder cataloging system is proposed based on a renewed method for estimating calcium value from a radiological image of the bone. A deliberate method was employed, the binning technique, for the input image which divides the input image into non-overlapping blocks to obtain accurate calcium volume estimation. In this proposed approach, the input image undergoes two stages of the process. In stage 1, input image preprocessing is accomplished with median filtering to eliminate the unwanted noise and it increases the quality of the image. Further, the processed image is fed to the Otsu-thresholding-segmentation method to highlight the affected regions from the processed bone image. The LBP (Local Binary Pattern) is a technique implemented to pull out the feature vector alone from the input image. Calcium value is estimated from abnormal regions from the segmented bone image and with the help of extracted texture features, the calcium concentration is obtained. MSVM (Multi-class Support Vector Machine) technique is applied to categorize as normal, osteoporosis, and osteopenia. In stage 2, the entire input is divided into 4 x 4 bins and preprocessing, segmentation, feature extraction, and calcium estimation process were applied similar to stage I to each bin separately and the calcium values of all bins are added together. Finally, stage 1 and stage 2 calcium values are summed up to obtain a more precise calcium estimation of the input image the feature vectors which were pull-out from others. The result can prove that the proposed binning technique is best for the bone disorder classification system which attained the greater accuracy of 97.4% and sensitivity of 98.3% when compared with and without binning technique. Validation of the results was performed with bone images, and these bone images were declared by the physician as bone disorder-affected images. The success rate of the bone disorder prediction is 80%.

Classifying Surface Points Based on Developability Using Machine Learning

The classifiers K-nearest neighbor (KNN), Multiclass support vector machine (MSVM), Decision Tree (DT), Discriminate Analysis (DA), Naive Bayes (NB), Random Forest (RF), and Ensemble Tree (ET) are the most well-known methods in machine learning. They are used in many fields like pattern recognition, medical disease analysis, user smartphone classification, text classification, etc. This paper presents a new framework for 3D surface point type classification using the most known classification methods in machine learning and the principal curvatures, the binormal vector, the cosine value of the angle between the normal vector and binormal vectors. The purpose of this study is to classify data points according to their developability. Also, the comparison between these methods is given to measure developability based on the accuracy and the processing time using several 3D surface examples.

A Novel Methodology for Disease Identification Using Metaheuristic Algorithm and Aura Image

Every human has a specific Aura. Every organism in the human body emits energy comprising of Ultra Violet radiation, thermal radiation, and electromagnetic radiation. These energy levels help to underline the physical health inside the human body. In general, these energy levels are called Aura. In order to capture the energy levels, specific cameras like Kirlian are used. These cameras try to capture the energy distribution and map them to the individual organs of the human body. In this article, we present a methodology using Image processing techniques, where Bivariate Gaussian Mixture Model (BGMM) is considered as a classifier to identify the diseases in humans based on the energy distribution. In this article, we have considered five different categories of diseased organs that are identified based on the energy distribution. The preprocessing is subjected to the morphological technique and Particle Swarm Optimization (PSO) algorithm is considered for feature extraction. The segmentation process is carried out using the feature extracted and training is carried out using the BGMM classifier. The result obtained is summarized using various other methods like Support Vector Machine (SVM), Artificial Neural Network (ANN), and Multiclass SVM (MSVM). The results showcase that the proposed methodology exhibits recognition accuracy at 90%.