Deep Feature Selection Research Articles

A Two-Stage Deep Feature Selection Method for Online Handwritten Bangla and Devanagari Basic Character Recognition

A Two-Stage Deep Feature Selection Method for Online Handwritten Bangla and Devanagari Basic Character Recognition

Multi-Class Classification of Breast Cancer Using 6B-Net with Deep Feature Fusion and Selection Method.

Breast cancer has now overtaken lung cancer as the world’s most commonly diagnosed cancer, with thousands of new cases per year. Early detection and classification of breast cancer are necessary to overcome the death rate. Recently, many deep learning-based studies have been proposed for automatic diagnosis and classification of this deadly disease, using histopathology images. This study proposed a novel solution for multi-class breast cancer classification from histopathology images using deep learning. For this purpose, a novel 6B-Net deep CNN model, with feature fusion and selection mechanism, was developed for multi-class breast cancer classification. For the evaluation of the proposed method, two large, publicly available datasets, namely, BreaKHis, with eight classes containing 7909 images, and a breast cancer histopathology dataset, containing 3771 images of four classes, were used. The proposed method achieves a multi-class average accuracy of 94.20%, with a classification training time of 226 s in four classes of breast cancer, and a multi-class average accuracy of 90.10%, with a classification training time of 147 s in eight classes of breast cancer. The experimental outcomes show that the proposed method achieves the highest multi-class average accuracy for breast cancer classification, and hence, the proposed method can effectively be applied for early detection and classification of breast cancer to assist the pathologists in early and accurate diagnosis of breast cancer.

Deep unsupervised feature selection by discarding nuisance and correlated features

Modern datasets often contain large subsets of correlated features and nuisance features, which are not or loosely related to the main underlying structures of the data. Nuisance features can be identified using the Laplacian score criterion, which evaluates the importance of a given feature via its consistency with the Graph Laplacians’ leading eigenvectors. We demonstrate that in the presence of large numbers of nuisance features, the Laplacian must be computed on the subset of selected features rather than on the complete feature set. To do this, we propose a fully differentiable approach for unsupervised feature selection, utilizing the Laplacian score criterion to avoid the selection of nuisance features. We employ an autoencoder architecture to cope with correlated features, trained to reconstruct the data from the subset of selected features. Building on the recently proposed concrete layer that allows controlling for the number of selected features via architectural design, simplifying the optimization process. Experimenting on several real-world datasets, we demonstrate that our proposed approach outperforms similar approaches designed to avoid only correlated or nuisance features, but not both. Several state-of-the-art clustering results are reported. Our code is publically available at https://github.com/jsvir/lscae.

Comprehensive Study for Breast Cancer Using Deep Learning and Traditional Machine Learning

Comprehensive Study for Breast Cancer Using Deep Learning and Traditional Machine Learning

Efficient palmprint biometric identification systems using deep learning and feature selection methods

Over the past two decades, several studies have paid great attention to biometric palmprint recognition. Recently, most methods in literature adopted deep learning due to their high recognition accuracy and the capability to adapt with different acquisition palmprint images. However, high-dimensional data with a large number of uncorrelated and redundant features remain a challenge due to computational complexity issues. Feature selection is a process of selecting a subset of relevant features, which aims to decrease the dimensionality, reduce the running time, and improve the accuracy. In this paper, we propose efficient unimodal and multimodal biometric systems based on deep learning and feature selection. Our approach called simplified PalmNet–Gabor concentrates on the improvement of the PalmNet for fast recognition of multispectral and contactless palmprint images. Therefore, we used Log-Gabor filters in the preprocessing to increase the contrast of palmprint features. Then, we reduced the number of features using feature selection and dimensionality reduction procedures. For the multimodal system, we fused modalities at the matching score level to improve system performance. The proposed method effectively improves the accuracy of the PalmNet and reduces the number of features as well the computational time. We validated the proposed method on four public palmprint databases, two multispectral databases, CASIA and PolyU, and two contactless databases, Tongji and PolyU 2D/3D. Experiments show that our approach achieves a high recognition rate while using a substantially lower number of features.

Classifying the multi-omics data of gastric cancer using a deep feature selection method

Gastric cancer has the highest incidence among all types of malignant tumors. The rapid development of high-throughput gene technology has greatly promoted people’s understanding of gastric cancer at the molecular level. However, there is a lack of information in single omics data, so dimensionality reduction is an effective method to overcome the dimensionality disaster of omics data. omics data has the characteristics of being multivariate and high-dimensional, which affects the efficiency of classification. Therefore, dimensionality reduction is an effective method to overcome the dimensionality disaster of omics data. However, neural network learning algorithm is seldom used to improve classification accuracy when feature selection of multi-omics data is carried out, therefore, in this study, a random forest deep feature selection (RDFS) algorithm was proposed. By integrating gene expression (Exp) data and copy number variation (CNV) data, the dimensions of multi-omics data were reduced and improve the classification accuracy by using a random forest and deep neural network. The results showed that the accuracy and area under the curve (AUC) of multi-omics data were better than that of single-omics data under the RDFS algorithm. With other feature selection algorithms, RDFS also had a higher prediction accuracy and AUC. We also validated the effect of feature selection on RDFS. Finally, survival analysis was used to evaluate the important genes identified during feature selection and to obtain enrichment gene ontology (GO) terms and biological pathways for these genes.

A composite framework of deep multiple view human joints feature extraction and selection strategy with hybrid adaptive sunflower optimization‐whale optimization algorithm for human action recognition in video sequences

AbstractIn computer vision and pattern recognition field, video‐based human action recognition (HAR) is the most predominant research area. Object recognition is needed to recognize the subjects regarding video contents, which allows reactive enquiry in a large number of camera contents, mainly in security based platforms where there is a prevalent growth of closed circuit television cameras. Generally, object detectors that have high performance are trained on a large collection of public benchmarks. Identifying human activities from unconstrained videos is the primary challenging task. Further, the feature extraction and feature selection from these unconstrained videos is also considered as a challenging issue. For that, in this article a new composite framework of HAR model is constructed by introducing an efficient feature extraction and selection strategy. The proposed feature extraction model extracts multiple view features, human joints features based on the domain knowledge of the action and fuses them with deep high level features extracted by an improved fully resolution convolutional neural networks. Also, it optimizes the feature selection strategy using the hybrid whale optimization algorithm and adaptive sun flower optimization that maximizes the feature entropy, correlation. It minimizes the error rate for improving the recognition accuracy of the proposed composite framework. The proposed model is validated on four different datasets, namely, Olympics sports, Virat Release 2.0, HMDB51, and UCF 50 sports action dataset to prove its effectiveness. The simulation results show that the proposed composite framework outperforms all the existing human recognition model in terms of classification accuracy and detection rate.

Gaussian Optimized Deep Learning-based Belief Classification Model for Breast Cancer Detection

With the rapid increase of new cases with an increased mortality rate, cancer is considered the second and most deadly disease globally. Breast cancer is the most widely affected cancer worldwide, with an increased death rate percentage. Due to radiologists’ processing of mammogram images, many computer-aided diagnoses have been developed to detect breast cancer. Early detection of breast cancer will reduce the death rate worldwide. The early diagnosis of breast cancer using the developed computer-aided diagnosis (CAD) systems still needed to be enhanced by incorporating innovative deep learning technologies to improve the accuracy and sensitivity of the detection system with a reduced false positive rate. This paper proposed an efficient and optimized deep learning-based feature selection approach with this consideration. This model selects the relevant features from the mammogram images that can improve the accuracy of malignant detection and reduce the false alarm rate. Transfer learning is used in the extraction of features initially. Na ext, a convolution neural network, is used to extract the features. The two feature vectors are fused and optimized with enhanced Butterfly Optimization with Gaussian function (TL-CNN-EBOG) to select the final most relevant features. The optimized features are applied to the classifier called Deep belief network (DBN) to classify the benign and malignant images. The feature extraction and classification process used two datasets, breast, and MIAS. Compared to the existing methods, the optimized deep learning-based model secured 98.6% of improved accuracy on the breast dataset and 98.85% of improved accuracy on the MIAS dataset.

A Multilevel Deep Feature Selection Framework for Diabetic Retinopathy Image Classification

Diabetes or Diabetes Mellitus (DM) is the upset that happens due to high glucose level within the body. With the passage of time, this polygenic disease creates eye deficiency referred to as Diabetic Retinopathy (DR) which can cause a major loss of vision. The symptoms typically originate within the retinal space square in the form of enlarged veins, liquid dribble, exudates, haemorrhages and small scale aneurysms. In current therapeutic science, pictures are the key device for an exact finding of patients’ illness. Meanwhile, an assessment of new medicinal symbolisms stays complex. Recently, Computer Vision (CV) with deep neural networks can train models with high accuracy. The thought behind this paper is to propose a computerized learning model to distinguish the key precursors of Dimensionality Reduction (DR). The proposed deep learning framework utilizes the strength of selected models (VGG and Inception V3) by fusing the extracated features. To select the most discriminant features from a pool of features, an entropy concept is employed before the classification step. The deep learning models are fit for measuring the highlights as veins, liquid dribble, exudates, haemorrhages and miniaturized scale aneurysms into various classes. The model will ascertain the loads, which give the seriousness level of the patient’s eye. The model will be useful to distinguish the correct class of seriousness of diabetic retinopathy pictures.

Deep Feature Selection for Anomaly Detection Based on Pretrained Network and Gaussian Discriminative Analysis

Deep learning neural network serves as a powerful tool for visual anomaly detection and fault diagnosis, attributed to its strong abstractive interpretation ability in the representation domain. The deep features from neural networks that are pre-trained on the ImageNet classification task have been proved to be useful for anomaly detection based on Gaussian discriminant analysis. However, with the ever-increasing complexity of deep learning neural networks, the set of deep features becomes massive where redundancy appears to be inevitable. The redundant features increase computational cost and degrade the performance of the anomaly detection method. In this paper, we discuss the deep feature selection for the anomaly detection task and show how to reduce the redundancy in the representation domain. We propose a horizontal selection (dimensional reduction) method of features with subspace decomposition and a vertical selection to identify the most effective network layer for anomaly detection and fault diagnosis. We test the proposed method on two public datasets, one for anomaly detection task and one for fault diagnosis of bearings. We show the significance of different network layers and feature subspaces on anomaly detection tasks, and prove the effectiveness of the feature selection strategy.

Radar Emitter Recognition Based on Deep Feature Selection Network

Radar Emitter Recognition Based on Deep Feature Selection Network

Radar Emitter Recognition Based on Deep Feature Selection Network

Radar Emitter Recognition Based on Deep Feature Selection Network

Hybrid Deep Learning Model for Endoscopic Lesion Detection and Classification Using Endoscopy Videos.

In medical imaging, the detection and classification of stomach diseases are challenging due to the resemblance of different symptoms, image contrast, and complex background. Computer-aided diagnosis (CAD) plays a vital role in the medical imaging field, allowing accurate results to be obtained in minimal time. This article proposes a new hybrid method to detect and classify stomach diseases using endoscopy videos. The proposed methodology comprises seven significant steps: data acquisition, preprocessing of data, transfer learning of deep models, feature extraction, feature selection, hybridization, and classification. We selected two different CNN models (VGG19 and Alexnet) to extract features. We applied transfer learning techniques before using them as feature extractors. We used a genetic algorithm (GA) in feature selection, due to its adaptive nature. We fused selected features of both models using a serial-based approach. Finally, the best features were provided to multiple machine learning classifiers for detection and classification. The proposed approach was evaluated on a personally collected dataset of five classes, including gastritis, ulcer, esophagitis, bleeding, and healthy. We observed that the proposed technique performed superbly on Cubic SVM with 99.8% accuracy. For the authenticity of the proposed technique, we considered these statistical measures: classification accuracy, recall, precision, False Negative Rate (FNR), Area Under the Curve (AUC), and time. In addition, we provided a fair state-of-the-art comparison of our proposed technique with existing techniques that proves its worthiness.

DEEPSEL: A novel feature selection for early identification of malware in mobile applications

Smartphone applications have gained popularity in recent years due to the large footprint of mobile phone usage and availability of a large number of value-added applications. The official app stores (google, IOS, Microsoft, Amazon) provide a platform for hosting, publishing, distributing, and managing the mobile applications developed by companies and individuals. This mobile application ecosystem could be used to distribute the malicious apps which are specifically designed to track behavior of users, spy on the activities of users, and could be a threat to the privacy, confidentiality, and integrity of the users. In this paper, we present a novel approach called DEEPSEL (Deep Feature Selection), a deep learning-based method for the identification of malware and malicious codes within android applications. DEEPSEL uses a set of features to characterize the behavior of android applications and classify them as legitimate and malicious. The main contribution is characterized by the usage of particle swarm optimization for performing feature selection. We evaluated our approach on a public malware data-set which is composed of samples collected from 39 unique malware families. Our results show that the proposed method can achieve very good results with an accuracy of around 83.6% and an F-measure of around 82.5%.

A Hybrid Lightweight 1D CNN-LSTM Architecture for Automated ECG Beat-Wise Classification

In this paper we have utilized a hybrid lightweight 1D deep learning model that combines convolutional neural network (CNN) and long short-term memory (LSTM) methods for accurate, fast, and automated beat-wise ECG classification. The CNN and LSTM models were designed separately to compare with the hybrid CNN-LSTM model in terms of accuracy, number of parameters, and the time required for classification. The hybrid CNN-LSTM system provides an automated deep feature extraction and classification for six ECG beats classes including Normal Sinus Rhythm (NSR), atrial fibrillation (AFIB), atrial flutter (AFL), atrial premature beat (APB), left bundle branch block (LBBB), and right bundle branch block (RBBB). The hybrid model uses the CNN blocks for deep feature extraction and selection from the ECG beat. While the LSTM layer will learn how to extract contextual time information. The results show that the proposed hybrid CNN-LSTM model achieves high accuracy and sensitivity of 98.22% and 98.23% respectively. This model is light and fast in classifying ECG beats and superior to other previously used models which makes it very suitable for embedded systems designs that can be used in clinical applications for monitoring heart diseases in faster and more efficient manner.

A New Transferable Fault Diagnosis Approach of Rotating Machinery Based on Deep Autoencoder and Dominant Features Selection under Different Operating Conditions

In the actual industrial scenarios, most existing fault diagnosis approaches are faced with two challenges, insufficient labeled training data and distribution divergences between training and testing datasets. For the above issues, a new transferable fault diagnosis approach of rotating machinery based on deep autoencoder and dominant features selection is proposed in this article. First, maximal overlap discrete wavelet packet transform is applied for signals processing and mix-domains statistical feature extraction. Second, dominant features selection by importance score and differences between domains is proposed to select dominant features with high fault-discriminative ability and domain invariance. Then, selected dominant features are used for pretraining deep autoencoder (source model), which helps in enhancing the fault representative ability of deep features. The parameters of the source model are transferred to the target model, and normal state features from target domain are adopted for fine-tuning the target model. Finally, the target model is applied for fault patterns classification. Motor and bearing fault datasets are used for a series of experiments, and the results verify that the proposed methods have better cross-domain diagnosis performance than comparative models.

Pneumonia detection from lung X‐ray images using local search aided sine cosine algorithm based deep feature selection method

Pneumonia is a major cause of death among children below the age of 5 years, globally. It is especially prevalent in developing and underdeveloped nations where the risk factors for the disease such as unhygienic living conditions, high levels of pollution and overcrowding are higher. Radiological examination (usually X-ray scans) is conducted to detect pneumonia, yet it is prone to subjective variability and can lead to disagreements among different radiologists. To detect traces of pneumonia from X-ray images, a more robust method is therefore required, which can be achieved by using a computer-aided diagnosis (CAD) system. In this study, we develop a two-stage framework, using the combination of deep learning and optimization algorithms, which is both accurate and time-efficient. In its first stage, the proposed framework extracts feature using a customized deep learning model called DenseNet-201 following the concept of transfer learning to cope with the scanty available data. In the second stage, we then reduce the feature dimension using an improved sine cosine algorithm equipped with adaptive beta hill climbing-based local search algorithm. The optimized feature subset is utilized for the classification of “Pneumonia” and “Normal” X-ray images using a support vector machines classifier. Upon an evaluation on a publicly available data set, the proposed method demonstrates the highest accuracy of 98.36% and sensitivity of 98.79% with a feature reduction of 85.55% (74 features selected out of 512), using a five-fold cross-validation scheme. Extensive additional experiments on continuous benchmark functions as well as the CEC-2017 test suite further showcase the superiority and suitability of our proposed approach in application to real-valued optimization problems. The relevant codes for the proposed method can be found in https://github.com/soumitri2001/Pneumonia-Detection-Local-Search-aided-SCA.

A novel approach based on integration of convolutional neural networks and deep feature selection for short-term solar radiation forecasting

In this study, a novel deep solar forecasting approach is proposed based on the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), continuous wavelet transform (CWT), feature extraction networks, RReliefF feature selection, and extreme learning machine (ELM). The global solar radiation is decomposed into mode functions with the CEEMDAN method. The CWT reconstructs one-dimensional data into two-dimensional scalogram images to include both frequency and the time of the daily and hourly correlations. For the feature extraction process, a cascade convolutional neural network architecture, which consists of AlexNet and GoogLeNet, was designed to extract distinctive deep features. As the high-performance features provide a high level of forecasting accuracy, these are concatenated as the subset feature vector and RReliefF utilized to rank and select the most distinctive features from the subset. The designed ELM is then trained with the selected features and the fully-trained ELM model is then used to evaluate the forecast performance. In the experiments, root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) of the proposed method were observed as 0.0642, 0.0241, and 0.1201 for one-step ahead, 0.0686, 0.0285, and 0.1279 for two-step ahead, and 0.0724, 0.0315, and 0.1317 for three-step ahead, respectively. The obtained results show that the proposed method exhibits accurate and robust forecasting performance and outperforms conventional regression models.

Automated detection of bioimages using novel deep feature fusion algorithm and effective high-dimensional feature selection approach

The classification of bioimages plays an important role in several biological studies, such as subcellular localisation, phenotype identification and other types of histopathological examinations. The objective of the present study was to develop a computer-aided bioimage classification method for the classification of bioimages across nine diverse benchmark datasets. A novel algorithm was developed, which systematically fused the features extracted from nine different convolution neural network architectures. A systematic fusion of features boosts the performance of a classifier but at the cost of the high dimensionality of the fused feature set. Therefore, non-discriminatory and redundant features need to be removed from a high-dimensional fused feature set to improve the classification performance and reduce the time complexity. To achieve this aim, a method based on analysis of variance and evolutionary feature selection was developed to select an optimal set of discriminatory features from the fused feature set. The proposed method was evaluated on nine different benchmark datasets. The experimental results showed that the proposed method achieved superior performance, with a significant reduction in the dimensionality of the fused feature set for most bioimage datasets. The performance of the proposed feature selection method was better than that of some of the most recent and classical methods used for feature selection. Thus, the proposed method was desirable because of its superior performance and high compression ratio, which significantly reduced the computational complexity.

Hash Code Generation using Deep Feature Selection Guided Siamese Network for Content-Based Medical Image Retrieval

It is very pleasing for human health that medical knowledge has increased and the technological infrastructure improves medical systems. The widespread use of medical imaging devices has been instrumental in saving lives by allowing early diagnosis of many diseases. These medical images are stored in large databases for many purposes. These datasets are used when a suspicious diagnostic case is encountered or to gain experience for inexperienced radiologists. To fulfill these tasks, images similar to one query image are searched from within the large dataset. Accuracy and speed are vital for this process, which is called content-based image retrieval (CBIR). In the literature, the best way to perform a CBIR system is by using hash codes. This study provides an effective hash code generation method based on feature selection-based downsampling of deep features extracted from medical images. Firstly, pre-hash codes of 256-bit length for each image are generated using a pairwise siamese network architecture that works based on the similarity of two images. Having a pre-hash code between -1 and 1 makes it very easy to generate hash code in hashing algorithms. For this reason, all activation functions of the proposed convolutional neural network (CNN) architecture are selected as hyperbolic tanh. Finally, neighborhood component analysis (NCA) feature selection methods are used to convert pre-hash code to binary hash code. This also downsamples the hash code length to 32-bit, 64-bit, or 96-bit levels. The performance of the proposed method is evaluated using NEMA MRI and NEMA CT datasets.