Reduced Feature Vector Research Articles

Exploring Feature Selection and Classification Techniques to Improve the Performance of an Electroencephalography-Based Motor Imagery Brain-Computer Interface System.

The accuracy of classifying motor imagery (MI) activities is a significant challenge when using brain-computer interfaces (BCIs). BCIs allow people with motor impairments to control external devices directly with their brains using electroencephalogram (EEG) patterns that translate brain activity into control signals. Many researchers have been working to develop MI-based BCI recognition systems using various time-frequency feature extraction and classification approaches. However, the existing systems still face challenges in achieving satisfactory performance due to large amount of non-discriminative and ineffective features. To get around these problems, we suggested a multiband decomposition-based feature extraction and classification method that works well, along with a strong feature selection method for MI tasks. Our method starts by splitting the preprocessed EEG signal into four sub-bands. In each sub-band, we then used a common spatial pattern (CSP) technique to pull out narrowband-oriented useful features, which gives us a high-dimensional feature vector. Subsequently, we utilized an effective feature selection method, Relief-F, which reduces the dimensionality of the final features. Finally, incorporating advanced classification techniques, we classified the final reduced feature vector. To evaluate the proposed model, we used the three different EEG-based MI benchmark datasets, and our proposed model achieved better performance accuracy than existing systems. Our model's strong points include its ability to effectively reduce feature dimensionality and improve classification accuracy through advanced feature extraction and selection methods.

SEMANTIC SEGMENTATION AND CONTENT-BASED RETRIEVAL IN MULTIMEDIA IMAGE DATABASES

Brain tumors, particularly gliomas, pose a significant threat to global health, necessitating accurate and efficient diagnostic methods. Magnetic Resonance Imaging (MRI) serves as a crucial tool for diagnosing glioma grades, but interpretation is subject to variability, hindering treatment planning. Intra and inter-observer variability in radiological image interpretation impede effective therapeutic strategies for brain tumor patients. Accessing relevant images from vast medical databases for comparison and treatment planning is cumbersome and time-consuming. This paper proposes a Content-Based Medical Image Retrieval (CBIR) system utilizing Convolutional Neural Network (CNN)-based feature extraction, specifically employing the AlexNet architecture. The system employs KNN clustering for indexing the feature map database and implements Gain-based feature selection to reduce feature vector dimensionality. The proposed system underwent evaluation using BraTS 2018 and 2020 datasets with five-fold cross-validation. Achieving state-of-the-art performance, the system demonstrated a mean Average Precision of 98% and Precision of 97%, showcasing its efficacy in accurately retrieving similar pathological MRI brain images.

Measurement of the Compressive Strength of Concrete Using Modeling of Deep Hybrid Forest Regression

The paper proposes a deep hybrid forest regression-based modeling method for measuring the compressive strength (CS) of concrete. Then, the reduced feature vector is used as input to train multiple subforest models (SFM), the predicted values are selected from multiple subforests via the KNN (K-nearest neighbor) method to combine them to obtain the layer regression vector (LRV), and it is combined with the reduced feature vector to obtain the improved LRV, then the output of this layer is taken; second, the regression vector (RV) of the input layer enhancement layer is used as input to obtain the output of the second layer FM, and the steps are repeated until the output of the input layer FM is complete. Finally, the output of the FM of the first layer is obtained. Several SFMs are trained and the result is obtained. The final prognosis is obtained by arithmetically averaging the forecast results of the SFMs of this layer.

A Hybridized Feature Extraction Model for Offline Yorùbá Document Recognition

Document recognition is required to convert handwritten and text documents into digital equivalents, making them more easily accessible and convenient to store. This study combined feature extraction techniques for recognizing Yorùbá documents in an effort to preserve the cultural values and heritages of the Yorùbá people. Ten Yorùbá documents were acquired from Kwara State University’s Library, and ten indigenous literate writers wrote the handwritten version of the documents. These were digitized using HP Scanjet300 and pre-processed. The pre-processed image served as input to the Local Binary Pattern, Speeded-Up-Robust-Features and Histogram of Gradient. The combined extracted feature vectors were input into the Genetic Algorithm. The reduced feature vector was fed into Support Vector Machine. A 10-folds cross-validation was used to train the model: LBP-GA, SURF-GA, HOG-GA, LBP-SURF-GA, HOG-SURF-GA, LBP-HOG-GA and LBP-HOG-SURF-GA. LBP-HOG-SURF-GA for Yorùbá printed text gave 90.0% precision, 90.3% accuracy and 15.5% FPR. LBP-HOG-SURF-GA for Handwritten Yorùbá document showed 80.9% precision, 82.6% accuracy and 20.4% (FPR) LBP-HOG-SURF-GA for CEDAR gave 98.0% precision, 98.4% accuracy and 2.6% FPR. LBP-HOG-SURF-GA for MNIST gave 99% precision, 99.5% accuracy, 99.0% and 1.1% FPR. The results of the hybridized feature extractions (LBP-HOG-SURF) demonstrated that the proposed work improves significantly on the various classification metrics.

Leukocytes Classification for Leukemia Detection Using Quantum Inspired Deep Feature Selection.

Leukocytes, also referred to as white blood cells (WBCs), are a crucial component of the human immune system. Abnormal proliferation of leukocytes in the bone marrow leads to leukemia, a fatal blood cancer. Classification of various subtypes of WBCs is an important step in the diagnosis of leukemia. The method of automated classification of WBCs using deep convolutional neural networks is promising to achieve a significant level of accuracy, but suffers from high computational costs due to very large feature sets. Dimensionality reduction through intelligent feature selection is essential to improve the model performance with reduced computational complexity. This work proposed an improved pipeline for subtype classification of WBCs that relies on transfer learning for feature extraction using deep neural networks, followed by a wrapper feature selection approach based on a customized quantum-inspired evolutionary algorithm (QIEA). This algorithm, inspired by the principles of quantum physics, outperforms classical evolutionary algorithms in the exploration of search space. The reduced feature vector obtained from QIEA was then classified with multiple baseline classifiers. In order to validate the proposed methodology, a public dataset of 5000 images of five subtypes of WBCs was used. The proposed system achieves a classification accuracy of about 99% with a reduction of 90% in the size of the feature vector. The proposed feature selection method also shows a better convergence performance as compared to the classical genetic algorithm and a comparable performance to several existing works.

Improving the robustness and stability of a machine learning model for breast cancer prognosis through the use of multi-modal classifiers

Breast cancer is a deadly disease with a high mortality rate among PAN cancers. The advancements in biomedical information retrieval techniques have been beneficial in developing early prognosis and diagnosis systems for cancer patients. These systems provide the oncologist with plenty of information from several modalities to make the correct and feasible treatment plan for breast cancer patients and protect them from unnecessary therapies and their toxic side effects. The cancer patient’s related information can be collected using various modalities like clinical, copy number variation, DNA-methylation, microRNA sequencing, gene expression, and histopathological whole slide images. High dimensionality and heterogeneity in these modalities demand the development of some intelligent systems to understand related features to the prognosis and diagnosis of diseases and make correct predictions. In this work, we have studied some end-to-end systems having two main components : (a) dimensionality reduction techniques applied to original features from different modalities and (b) classification techniques applied to the fusion of reduced feature vectors from different modalities for automatic predictions of breast cancer patients into two categories: short-time and long-time survivors. Principal component analysis (PCA) and variational auto-encoders (VAEs) are used as the dimensionality reduction techniques, followed by support vector machines (SVM) or random forest as the machine learning classifiers. The study utilizes raw, PCA, and VAE extracted features of the TCGA-BRCA dataset from six different modalities as input to the machine learning classifiers. We conclude this study by suggesting that adding more modalities to the classifiers provides complementary information to the classifier and increases the stability and robustness of the classifiers. In this study, the multimodal classifiers have not been validated on primary data prospectively.

White Blood Cells Classification Using Entropy-Controlled Deep Features Optimization.

White blood cells (WBCs) constitute an essential part of the human immune system. The correct identification of WBC subtypes is critical in the diagnosis of leukemia, a kind of blood cancer defined by the aberrant proliferation of malignant leukocytes in the bone marrow. The traditional approach of classifying WBCs, which involves the visual analysis of blood smear images, is labor-intensive and error-prone. Modern approaches based on deep convolutional neural networks provide significant results for this type of image categorization, but have high processing and implementation costs owing to very large feature sets. This paper presents an improved hybrid approach for efficient WBC subtype classification. First, optimum deep features are extracted from enhanced and segmented WBC images using transfer learning on pre-trained deep neural networks, i.e., DenseNet201 and Darknet53. The serially fused feature vector is then filtered using an entropy-controlled marine predator algorithm (ECMPA). This nature-inspired meta-heuristic optimization algorithm selects the most dominant features while discarding the weak ones. The reduced feature vector is classified with multiple baseline classifiers with various kernel settings. The proposed methodology is validated on a public dataset of 5000 synthetic images that correspond to five different subtypes of WBCs. The system achieves an overall average accuracy of 99.9% with more than 95% reduction in the size of the feature vector. The feature selection algorithm also demonstrates better convergence performance as compared to classical meta-heuristic algorithms. The proposed method also demonstrates a comparable performance with several existing works on WBC classification.

Two New Indices for Unsupervised Dimensionality Reduction of Hyperspectral Data

Hyperspectral imaging, with hundreds of spectral bands provide a rich data source for thematic mapping. However, the high dimensionality of these images is a challenge which is commonly treated by dimensionality reduction techniques. In this paper, a pixel-based hyperspectral dimensionality reduction technique is proposed in which the spectral signature curve (SSC) is divided into segments with equal lengths. Afterwards two new mathematical indices, which are based on integral and L2-norm, are proposed to transform each SSC segment into a new feature of the reduced feature vector. The proposed method considers the ordinance of SSC and, unlike PCA, can be applied individually to all pixels independently and simultaneously which means the method is applicable in parallel processing. The proposed data reduction technique was applied to two well-known agricultural hyperspectral scenes and was compared to wavelet and PCA data reduction techniques. The obtained results proved the efficiency of the proposed method from both classification accuracy and processing time aspects.

Machine learning augmented multi-sensor data fusion to detect aero engine fan rotor blade flutter

Abstract Flutter-induced fatigue failure investigation of the fan blades of aero-engines necessitates extensive testing. During engine ground testing, strain gauges on rotor fan blades and casing vibration sensors were employed to investigate structural dynamic aspects. The correlation between strain sensor signals and fan casing vibration signals allowed the diagnosis of fluttering fan blades. For automated flutter detection during engine development testing, a machine learning-augmented information fusion methodology was developed. The method analyses casing vibration signals by extracting time-domain statistical features, intrinsic mode function characteristics through empirical mode decomposition, and recurrence quantification features. Feature vectors obtained from a relatively large set of engine tests were subjected to dimension reduction by applying machine learning techniques to rank them. Reduced feature vector space was labelled as “flutter” or “normal” based on the correlation of rotor strain gauge signals. In addition, the labelled feature vectors were employed to train classifier models using supervised learning-based algorithms such as Support Vector Machines, Linear Discriminant Analysis, K-means Clustering, and Artificial Neural Networks. Using only vibration signals from the casing, the trained and validated classifiers were able to detect flutter in fan baldes with a 99% probability during subsequent testing.

Automated Schizophrenia detection using local descriptors with EEG signals

Schizophrenia (SZ) is a severe mental disorder characterized by behavioral imbalance and impaired cognitive ability. This paper proposes a local descriptors-based automated approach for SZ detection using electroencephalogram (EEG) signals. Specifically, we introduce a local descriptor, histogram of local variance (HLV), for feature representation of EEG signals. The HLV is generated by using locally computed variances. In addition to HLV, symmetrically weighted-local binary patterns (SLBP)-based histogram features are also computed from the multi-channel EEG signals. Thus, obtained HLV and SLBP-based features are given to a correlation-based feature selection algorithm to reduce the length of the feature vector. Finally, the reduced feature vector is fed to an AdaBoost classifier to classify SZ and healthy EEG signals. Besides, we have tested the influence of the different lobe regions in detecting SZ. For this, we combined the features extracted from channels belonging to the same group and performed the classification. Experimental results on two publicly available datasets suggest the local descriptors computed from temporal lobe channels are very effective in capturing regional variations of EEG signals. The proposed local-descriptors-based approach obtained an average classification accuracy of 92.85% and 99.36% on Dataset-1 and Dataset-2, respectively, with only a feature vector of length 13.

Fault Diagnosis of Coal Mill Based on Kernel Extreme Learning Machine with Variational Model Feature Extraction

Aiming at the typical faults in the coal mills operation process, the kernel extreme learning machine diagnosis model based on variational model feature extraction and kernel principal component analysis is offered. Firstly, the collected signals of vibration and loading force, corresponding to typical faults of coal mill, are decomposed by variational model decomposition, and the intrinsic model functions at different scales are obtained. Then, the eigenvectors consisting of feature energy and sample entropy in these functions are respectively calculated, and the kernel principal component analysis is used for noise removal and dimensionality reduction. Finally, the kernel extreme learning machine model is trained and tested with the dimension reduced feature vector as input and the corresponding coal mill state as output. The results show that the variational model decomposition extraction can improve the input features of the model compared with the single eigenvector model, and the kernel principal component analysis method can significantly reduce the information redundancy and the correlation of eigenvectors, which can effectively save time and cost, and improve the prediction performance of the model to some extent. The establishment of this model provides a new idea for the study of coal mill fault diagnosis.

Early prediction of student engagement in virtual learning environments using machine learning techniques

Learning analytics aims at helping the students to attain their learning goals. The predictions in learning analytics are made to enhance the effectiveness of educational interferences. This study predicts student engagement at an early phase of a Virtual Learning Environment (VLE) course by analyzing data collected from consecutive years. The prediction model is developed using machine learning techniques applied to a subset of Open University Learning Analytics Dataset, provided by Open University (OU), Britain. The investigated data belongs to 7,775 students who attended social science courses for consecutive assessment years. The experiments are conducted with a reduced feature set to predict whether the students are highly or lowly engaged in the courses. The attributes indicating students' interaction with the VLE, their scores, and final results are the most contributing variables for the predictive analysis. Based on these variables, a reduced feature vector is constructed. The baseline used in the study is the linear regression model. The model’s best results showed 95% accurate, 95% precise, and 98% relevant results with the Random Forest classification algorithm. Early prediction’s relevant features are a subset of click activities, which provided a functional interface between the students and the VLE.

A Dermoscopic Inspired System for Localization and Malignancy Classification of Melanocytic Lesions

This study aims at developing a clinically oriented automated diagnostic tool for distinguishing malignant melanocytic lesions from benign melanocytic nevi in diverse image databases. Due to the presence of artifacts, smooth lesion boundaries, and subtlety in diagnostic features, the accuracy of such systems gets hampered. Thus, the proposed framework improves the accuracy of melanoma detection by combining the clinical aspects of dermoscopy. Two methods have been adopted for achieving the aforementioned objective. Firstly, artifact removal and lesion localization are performed. In the second step, various clinically significant features such as shape, color, texture, and pigment network are detected. Features are further reduced by checking their individual significance (i.e., hypothesis testing). These reduced feature vectors are then classified using SVM classifier. Features specific to the domain have been used for this design as opposed to features of the abstract images. The domain knowledge of an expert gets enhanced by this methodology. The proposed approach is implemented on a multi-source dataset (PH2 + ISBI 2016 and 2017) of 515 annotated images, thereby resulting in sensitivity, specificity and accuracy of 83.8%, 88.3%, and 86%, respectively. The experimental results are promising, and can be applied to detect asymmetry, pigment network, colors, and texture of the lesions.

A Novel Deep-Learning-Based Framework for the Classification of Cardiac Arrhythmia.

Cardiovascular diseases (CVDs) are the primary cause of death. Every year, many people die due to heart attacks. The electrocardiogram (ECG) signal plays a vital role in diagnosing CVDs. ECG signals provide us with information about the heartbeat. ECGs can detect cardiac arrhythmia. In this article, a novel deep-learning-based approach is proposed to classify ECG signals as normal and into sixteen arrhythmia classes. The ECG signal is preprocessed and converted into a 2D signal using continuous wavelet transform (CWT). The time–frequency domain representation of the CWT is given to the deep convolutional neural network (D-CNN) with an attention block to extract the spatial features vector (SFV). The attention block is proposed to capture global features. For dimensionality reduction in SFV, a novel clump of features (CoF) framework is proposed. The k-fold cross-validation is applied to obtain the reduced feature vector (RFV), and the RFV is given to the classifier to classify the arrhythmia class. The proposed framework achieves 99.84% accuracy with 100% sensitivity and 99.6% specificity. The proposed algorithm outperforms the state-of-the-art accuracy, F1-score, and sensitivity techniques.

Mode classification for vortex shedding from an oscillating wind turbine using machine learning

This study presents an effective strategy that applies machine learning methods to classify vortex shedding modes produced by the oscillating cylinder of a bladeless wind turbine. A 2-dimensional computational fluid dynamic (CFD) simulation using OpenFOAMv2006 was developed to simulate a bladeless wind turbines vortex shedding behavior. The simulations were conducted at two wake modes (2S, 2P) and a transition mode (2PO). The local flow measurements were recorded using four sensors: vorticity, flow speed, stream-wise and transverse stream-wise velocity components. The time-series data was transformed into the frequency domain to generate a reduced feature vector. A variety of supervised machine learning models were quantitatively compared based on classification accuracy. The best performing models were then reevaluated based on the effects of artificial noisy experimental data on the models’ performance. The velocity sensors orientated transverse to the pre-dominant flow (u y ) achieved improved testing accuracy of 15% compared to the next best sensor. The random forest and k-nearest neighbor models, using u y , achieved 99.3% and 99.8% classification accuracy, respectively. The feature noise analysis conducted reduced classification accuracy by 11.7% and 21.2% at the highest noise level for the respective models. The random forest algorithm trained using the transverse stream-wise component of the velocity vector provided the best balance of testing accuracy and robustness to data corruption. The results highlight the proposed methods’ ability to accurately identify vortex structures in the wake of an oscillating cylinder using feature extraction.

Schizophrenia Detection in Adolescents from EEG Signals using Symmetrically weighted Local Binary Patterns.

Schizophrenia is one of the most complex of all mental diseases. In this paper, we propose a symmetrically weighted local binary patterns (SLBP)-based automated approach for detection of schizophrenia in adolescents from electroencephalogram (EEG) signals. We extract SLBP-based histogram features from each of the EEG channels. These features are given to a correlation-based feature selection algorithm to get reduced feature vector length. Finally, the feature vector thus obtained is given to LogitBoost classifier to discriminate between schizophrenia and healthy EEG signals.The results validated on the publicly available database suggest that the SLBP effectively characterize the changes in EEG signals and are helpful for the classification of schizophrenia and healthy EEG signals with a classification accuracy of 91.66%. In addition, our approach has provided better results than the recently proposed approaches in schizophrenia detection.

Novel architecture with selected feature vector for effective classification of mitotic and non-mitotic cells in breast cancer histology images

The paper focuses on the detection of mitosis in breast cancer. Detection methods in vogue rely heavily on visual inspection and assessment of histology images by trained pathologists, so the validation of results is a time-consuming process. Even accuracy wise, there remains room for improvement. This calls for high accuracy automated systems to replace manual assessments. Mitotic count is important because it provides us information about the aggressiveness of the tumour cells i.e. how rapidly it is spreading in the human body. To address this research gap, the current study is devoted to construct a novel framework utilizing neural network-based concepts along with reduced feature vectors and multiple machine learning techniques to classify the mitotic and non-mitotic cells. The dominant features in the detection of mitosis are texture features, therefore the texture of cells is used to derive the efficient reduced feature vectors which plays a significant role in identifying cancerous cells. During the feature engineering process, the features that distinguish a cancer cell from the normal cells were identified. Using these features as the base features, extraction with multiple techniques like GLCM, LBP and LTP along with classification algorithms like SVM, Naive Bayes and Random Forest were applied to the source dataset of the images, whose ground truth reference was available. Inspiring from Neural Networks, the proposed architecture allocates different weights to different features which increases the productivity of the model. The proposed framework is a new and intelligent version of ensemble learning which makes ensemble learning a more powerful tool for accurate and high-throughput results. The proposed method outperforms techniques already reported in the literature for the detection of mitotic cells in MITOS-12, AMIDA 13, MITOS-14 and TUPAC16.

A study on multi-class anxiety detection using wearable EEG headband

In this paper, we present a trait anxiety detection framework using resting-state electroencephalography (EEG) data. Our proposed framework consists of EEG data acquisition, pre-processing, feature extraction and selection, and classification stages. EEG data of 65 participants is recorded in an eye-open state for the duration of two minutes. The trait anxiety scores are gathered using the state-trait anxiety inventory questionnaire, which is used to label the participant’s EEG data into two (non-anxious and anxious) and three (non-anxious, low anxious, and highly anxious) classes. Pre-processing of the recorded EEG data is performed using the onboard noise cancellation scheme of the MUSE EEG headband. Channel selection is performed by applying a t-test and analysis of variance on the power spectral densities for two and three classes of anxiety respectively. Five time-domain features are extracted from the selected EEG channels. The wrapper method for feature selection is applied for selecting an optimum subset of features, which are used to classify the trait anxiety. The highest classification accuracy of $$87.69\%$$ and $$83.07\%$$ using random forest classifier is achieved for two and three class anxiety classification respectively. Our proposed trait anxiety detection scheme outperforms existing schemes in terms of higher classification accuracy and reduced feature vector length.

Automatic diagnosis of coronavirus (COVID-19) using shape and texture characteristics extracted from X-Ray and CT-Scan images

Automatic diagnosis of coronavirus (COVID-19) is studied in this research. Deep learning methods especially convolutional neural networks (CNNs) have shown great success in COVID-19 diagnosis in recent works. But they are efficient when the depth of network is high enough. However, the use of a deep network requires a sufficiently large training set, which is not available in practice. From the other hand, the use of a shallow CNN may not provide superior results because it is not able to rich feature extraction due to lacking enough convolutional layers. To deal with this difficulty, the contextual features reduced by convolutional filters (CFRCF) is proposed in this work. CFRCF extracts shape and textural features as contextual feature maps from the chest X-ray radiographs and abdominal computed tomography (CT) images. Morphological operators, Gabor filter banks and attribute filters are used for contextual feature extraction. Then, two convolutional filters are applied to the contextual feature cube to extract the nonlinear sub-features and hidden relationships among the contextual features. Finally, a fully connected layer is used to produce a reduced feature vector which is fed to a classifier. Support vector machine and random forest are used as classifier. The experimental results show the superior performance of the proposed method from the recognition accuracy and running time point of view using limited training samples. More than 76% and 94% overall classification accuracy is obtained by the proposed method in CT scan and X-ray images datasets, respectively.

A feature-level fusion based improved multimodal biometric recognition system using ear and profile face

Ear based identity recognition subject to uncontrolled conditions such as illumination changes, pose variation, low contrast, partial occlusion and noise, is an active research area in the field of biometrics. Meanwhile, multimodal biometrics is becoming increasingly popular and offers improved performance due to the use of multiple sources of information. In this paper, a multimodal biometrics system is proposed based on the ear and profile face that not only alleviates the short-comings of ear biometrics but also improves the overall recognition rate. The ear and profile face modalities are first represented individually using the combination of two efficient local feature descriptors namely, local phase quantization (LPQ) and local directional patterns (LDP). These histogram-based local descriptors are then combined into a high-dimensional feature vector that preserves complementary information in both frequency and spatial domains. The PCA along with the z-score normalization technique is independently applied on each feature vector and the resultant reduced feature vectors are combined at the feature level. The kernel discriminative common vector (KDCV) approach is finally exploited over the combined feature set to derive more discriminative and non-linear features for the identification of individuals using kNN classifier. The effectiveness of the proposed model has been verified with the deep features derived from three popular pre-trained CNN models such as AlexNet, VGG16 and GoogleNet. Experimental results on two benchmark databases clearly show that the proposed approach achieves better performance than individual modality and other state-of-the-art methods.