Search

Deep-learning models play a significant role in modern software solutions, with the capabilities of handling complex tasks, improving accuracy, automating processes, and adapting to diverse domains, eventually contributing to advancements in various industries. This study provides a comparative study on deep-learning techniques that can also be deployed on resource-constrained edge devices. As a novel contribution, we analyze the performance of seven Convolutional Neural Network models in the context of data augmentation, feature extraction, and model compression using acoustic data. The results show that the best performers can achieve an optimal trade-off between model accuracy and size when compressed with weight and filter pruning followed by 8-bit quantization. In adherence to the study workflow utilizing the forest sound dataset, MobileNet-v3-small and ACDNet achieved accuracies of 87.95% and 85.64%, respectively, while maintaining compact sizes of 243 KB and 484 KB, respectively. Henceforth, this study concludes that CNNs can be optimized and compressed to be deployed in resource-constrained edge devices for classifying forest environment sounds. [ABSTRACT FROM AUTHOR]

The existing methods for monitoring vessels are mainly based on radar and automatic identification systems. Additional sensors that are used include video cameras. Such systems feature cameras that capture images and software that analyzes the selected video frames. Methods for the classification of non-conventional vessels are not widely known. These methods, based on image samples, can be considered difficult. This paper is intended to show an alternative way to approach image classification problems; not by classifying the entire input data, but smaller parts. The described solution is based on splitting the image of a ship into smaller parts and classifying them into vectors that can be identified as features using a convolutional neural network (CNN). This idea is a representation of a bag-of-words mechanism, where created feature vectors might be called words, and by using them a solution can assign images a specific class. As part of the experiment, the authors performed two tests. In the first, two classes were analyzed and the results obtained show great potential for application. In the second, the authors used much larger sets of images belonging to five vessel types. The proposed method indeed improved the results of classic approaches by 5%. The paper shows an alternative approach for the classification of non-conventional vessels to increase accuracy. [ABSTRACT FROM AUTHOR]

Deep learning has been successfully applied to classification tasks in many fields due to its good performance in learning discriminative features. However, the application of deep learning to printing defect classification is very rare, and there is almost no research on the classification method for printing defects with imbalanced samples. In this paper, we present a deep convolutional neural network model to extract deep features directly from printed image defects. Furthermore, considering the asymmetry in the number of different types of defect samples—that is, the number of different kinds of defect samples is unbalanced—seven types of over-sampling methods were investigated to determine the best method. To verify the practical applications of the proposed deep model and the effectiveness of the extracted features, a large dataset of printing detect samples was built. All samples were collected from practical printing products in the factory. The dataset includes a coarse-grained dataset with four types of printing samples and a fine-grained dataset with eleven types of printing samples. The experimental results show that the proposed deep model achieves a 96.86% classification accuracy rate on the coarse-grained dataset without adopting over-sampling, which is the highest accuracy compared to the well-known deep models based on transfer learning. Moreover, by adopting the proposed deep model combined with the SVM-SMOTE over-sampling method, the accuracy rate is improved by more than 20% in the fine-grained dataset compared to the method without over-sampling. [ABSTRACT FROM AUTHOR]

Electric load identification and classification for smart grid environment can improve the power service for both consumers and producers. The main concept of electric load identification and classification is to disaggregate various loads and categorize them. In this paper, a new practical method for electric load identification and classification is presented. The method is based on using a power monitor to analyze a real measured current waveform of a grid-connected device. A set number of features is extracted using the currents' physical components-based power theory decomposition. Using currents' physical components ensures a constant number of features, which maintains the signal's characteristics regardless of the harmonic content. These features are used to train a supervised classifier based on two techniques: artificial neural network and nearest neighbor search. The theory is outlined, and experimental results are shown. This paper demonstrates high accuracy performance in identifying an electric load from a designated database. Furthermore, the results show a definite classification of an untrained operation state of a device to the closest trained operation state, for example, the excitation angle of a dimmer. In a comparative study, the method is shown to outperform other state-of-the-art techniques, which are based on harmonic components. [ABSTRACT FROM AUTHOR]

In detecting large-scale attacks, deep neural networks (DNNs) are an effective approach based on high-quality training data samples. Feature selection and feature extraction are the primary approaches for data quality enhancement for high-accuracy intrusion detection. However, their enhancement root causes usually present weak relationships to the differences between normal and attack behaviors in the data samples. Thus, we propose a Classification Tendency Difference Index (CTDI) model for feature selection and extraction in intrusion detection. The CTDI model consists of three indexes: Classification Tendency Frequency Difference (CTFD), Classification Tendency Membership Difference (CTMD), and Classification Tendency Distance Difference (CTDD). In the dataset, each feature has many feature values (FVs). In each FV, the normal and attack samples indicate the FV classification tendency, and CTDI shows the classification tendency differences between the normal and attack samples. CTFD is the frequency difference between the normal and attack samples. By employing fuzzy C means (FCM) to establish the normal and attack clusters, CTMD is the membership difference between the clusters, and CTDD is the distance difference between the cluster centers. CTDI calculates the index score in each FV and summarizes the scores of all FVs in the feature as the feature score for each of the three indexes. CTDI adopts an Auto Encoder for feature extraction to generate new features from the dataset and calculate the three index scores for the new features. CTDI sorts the original and new features for each of the three indexes to select the best features. The selected CTDI features indicate the best classification tendency differences between normal and attack samples. The experiment results demonstrate that the CTDI features achieve better detection accuracy as classified by DNN for the Aegean WiFi Intrusion Dataset than their related works, and the detection enhancements are based on the improved classification tendency differences in the CTDI features. [ABSTRACT FROM AUTHOR]

This paper proposes a detection and classification method of recessive weakness in Superbuck converter through wavelet packet decomposition (WPD) and principal component analysis (PCA) combined with probabilistic neural network (PNN). The Superbuck converter presents excellent performance in many applications and is also faced with today's demands, such as higher reliability and steadier operation. In this paper, the detection and classification issue to recessive weakness is settled. Firstly, the performance of recessive weakness both in the time and frequency domain are demonstrated to clearly show the actual deterioration of the circuit system. The WPD and Parseval's theorem are utilized in this paper to feature the extraction of recessive weakness. The energy discrepancy of the fault signals at different wavelet decomposition levels are then chosen as the feature vectors. PCA is also employed to the dimensionality reduction of feature vectors. Then, a probabilistic neural network is applied to automatically detect and classify the recessive weakness from different components on the basis of the extracted features. Finally, the classification accuracy of the proposed classification algorithm is verified and tested with experiments, which present satisfying classification accuracy. [ABSTRACT FROM AUTHOR]

Abstract Features learned by deep Convolutional Neural Networks (CNNs) have been recognized to be more robust and expressive than hand-crafted ones. They have been successfully used in different computer vision tasks such as object detection, pattern recognition and image understanding. Given a CNN architecture and a training procedure, the efficacy of the learned features depends on the domain-representativeness of the training examples. In this paper we investigate the use of CNN-based features for the purpose of food recognition and retrieval. To this end, we first introduce the Food-475 database, that is the largest publicly available food database with 475 food classes and 247,636 images obtained by merging four publicly available food databases. We then define the food-domain representativeness of different food databases in terms of the total number of images, number of classes of the domain and number of examples for class. Different features are then extracted from a CNN based on the Residual Network with 50 layers architecture and trained on food databases with diverse food-domain representativeness. We evaluate these features for the tasks of food classification and retrieval. Results demonstrate that the features extracted from the Food-475 database outperform the other ones showing that we need larger food databases in order to tackle the challenges in food recognition, and that the created database is a step forward toward this end. Highlights • CNN-based features are used for the purpose of food classification and retrieval. • Several CNN architectures are compared and Residual Networks (ResNet-50) are the best. • Fine-tuned and trained ResNet-50 are used as feature extractors for food recognition. • A new benchmark food database which contains 475 food classes is introduced. • Our results show that the most robust features for food classification and retrieval are those obtained from the ResNet-50 fine-tuned on the Food-475 database. [ABSTRACT FROM AUTHOR]

Neural networks overcome drawbacks of vision tasks by becoming convolutional in a wide range of layers. The salient map is affected by multilevels of strong pixels (superpixels) in global images and that is dependent on the hard threshold for their dividing. Deep neural networks have been established for saliency prediction of segmentation because the feature extraction must be suited to the input data. The convolutional neural network (CNN) also endures conflict between spatial pattern and a likeness of salient objects. Semantic segmentation is one of the approaches to continue classification based on these features. Therefore, upgrading the extraction process can be of use in saliency. In this work, we optimize DeepLab based on an atrous convolutional and a conditional random field (CRF) with a bottleneck in the semantic segmentation method, which serves for classification. The backbone of deep feature extraction is atrous convolution and the bottleneck based on CRF for hybrid saliency in the encoder-decoder system. The classification results are compared with some approaches for saliency prediction of recent deeper methods in an ISIC 2017 dataset. The results give better values not only for saliency prediction for segmentation but also for training and testing for classification. [ABSTRACT FROM AUTHOR]

Spatial and spectral information are obtained simultaneously by hyperspectral remote sensing. Joint extraction of these information of hyperspectral image is one of most import methods for hyperspectral image classification. In this paper, a novel deep convolutional neural network (CNN) is proposed, which extracts spectral-spatial information of hyperspectral images correctly. The proposed model not only learns sufficient knowledge from the limited number of samples, but also has powerful generalization ability. The proposed framework based on three-dimensional convolution can extract spectral-spatial features of labeled samples effectively. Though CNN has shown its robustness to distortion, it cannot extract features of different scales through the traditional pooling layer that only have one size of pooling window. Hence, spatial pyramid pooling (SPP) is introduced into three-dimensional local convolutional filters for hyperspectral classification. Experimental results with a widely used hyperspectral remote sensing dataset show that the proposed model provides competitive performance. [ABSTRACT FROM AUTHOR]

A review on the recent segmentation and tumor grade classification techniques of brain Magnetic Resonance (MR) Images is the objective of this paper. The requisite for early detection of a brain tumor and its grade is the motivation for this study. In Magnetic Resonance Imaging (MRI), the tumor might appear clear but physicians need quantification of the tumor area for further treatment. This is where the digital image processing methodologies along with machine learning aid further diagnosis, treatment, prior and post-surgical procedures, synergizing between the radiologist and computer. These hybrid techniques provide a second opinion and assistance to radiologists in understanding medical images hence improving diagnostic accuracy. This article aims to retrospect the current trends in segmentation and classification relevant to tumor infected human brain MR images with a target on gliomas which include astrocytoma. The methodologies used for extraction and grading of tumors which can be integrated into the standard clinical imaging protocols are elucidated. Lastly, a crucial assessment of the state of the art, future developments and trends are dissertated. [ABSTRACT FROM AUTHOR]

The purpose of this work is to increase the security of the perimeter of an area from unauthorized intrusions by creating an improved algorithm for classifying acoustic impacts recorded with a sensor system based on a phase-sensitive optical time reflectometer (phi-OTDR). The algorithm includes machine learning, so a dataset consisting of two classes was assembled. The dataset consists of two classes. The first class is the data of the steps, and the second class is other non-stepping influences (engine noise, a passing car, a passing cyclist, etc.). As an intrusion signal, a human walking signal is analyzed and recorded in frames of 5 s, which passed the threshold condition. Since, in most cases, the intruder moves on foot to overcome the perimeter, the analysis of the acoustic effects generated during the step will increase the efficiency of the perimeter detection tools. When walking quietly, step signals can be quite weak, and background signals can contain high energy and visually resemble the signals you are looking for. Therefore, an algorithm was created that processes space–time diagrams developed in real time, which are grayscale images. At the same time, during the processing of one image, two more images are calculated, which are the result of processing the denoised autoencoder and the created mathematical model of the adaptive correlation. Then, the three obtained images are fed to the input of the created three-channel neural network classifier, which includes convolutional layers for the automatic extraction of spatial features. The probability of correctly detecting steps is 98.3% and that of background actions is 97.93%. [ABSTRACT FROM AUTHOR]

Person re-identification is the task of matching pedestrian images across a network of non-overlapping camera views. It poses aggregated challenges resulted from random human pose, clutter from the background, illumination variations, and other factors. There has been a vast number of studies in recent years with promising success. However, key challenges have not been adequately addressed and continue to result in sub-optimal performance. Attention-based person re-identification gains more popularity in identifying discriminatory features from person images. Its potential in terms of extracting features common to a pair of person images across the feature extraction pipeline has not been be fully exploited. In this paper, we propose a novel attention-based Siamese network driven by a mutual-attention module decomposed into spatial and channel components. The proposed mutual-attention module not only leads feature extraction to the discriminative part of individual images, but also fuses mutual features symmetrically across pairs of person images to get informative regions common to both input images. Our model simultaneously learns feature embedding for discriminative cues and the similarity measure. The proposed model is optimized with multi-task loss, namely classification and verification loss. It is further optimized by a learnable mutual-attention module to facilitate an efficient and adaptive learning. The proposed model is thoroughly evaluated on extensively used large-scale datasets, Market-1501 and Duke-MTMC-ReID. Our experimental results show competitive results with the state-of-the-art works and the effectiveness of the mutual-attention module. [ABSTRACT FROM AUTHOR]

The convolutional neural network (CNN) has been gradually applied to the hyperspectral images (HSIs) classification, but the lack of training samples caused by the difficulty of HSIs sample marking and ignoring of correlation between spatial and spectral information seriously restrict the HSIs classification accuracy. In an attempt to solve these problems, this paper proposes a dual-branch extraction and classification method under limited samples of hyperspectral images based on deep learning (DBECM). At first, a sample augmentation method based on local and global constraints in this model is designed to augment the limited training samples and balance the number of different class samples. Then spatial-spectral features are simultaneously extracted by the dual-branch spatial-spectral feature extraction method, which improves the utilization of HSIs data information. Finally, the extracted spatial-spectral feature fusion and classification are integrated into a unified network. The experimental results of two typical datasets show that the DBECM proposed in this paper has certain competitive advantages in classification accuracy compared with other public HSIs classification methods, especially in the Indian pines dataset. The parameters of the overall accuracy (OA), average accuracy (AA), and Kappa of the method proposed in this paper are at least 4.7%, 5.7%, and 5% higher than the existing methods. [ABSTRACT FROM AUTHOR]

3D convolutional neural networks (CNNs) have been demonstrated to be a powerful tool in hyperspectral images (HSIs) classification. However, using the conventional 3D CNNs to extract the spectral–spatial feature for HSIs results in too many parameters as HSIs have plenty of spatial redundancy. To address this issue, in this paper, we first design multiscale convolution to extract the contextual feature of different scales for HSIs and then propose to employ the octave 3D CNN which factorizes the mixed feature maps by their frequency to replace the normal 3D CNN in order to reduce the spatial redundancy and enlarge the receptive field. To further explore the discriminative features, a channel attention module and a spatial attention module are adopted to optimize the feature maps and improve the classification performance. The experiments on four hyperspectral image data sets demonstrate that the proposed method outperforms other state-of-the-art deep learning methods. [ABSTRACT FROM AUTHOR]

In recent years, Peripheral blood smear is a generic analysis to assess the person’s health status. Manual testing of Peripheral blood smear images are difficult, time-consuming and is subject to human intervention and visual error. This method encouraged for researchers to present algorithms and techniques to perform the peripheral blood smear analysis with the help of computer-assisted and decision-making techniques. Existing CAD based methods are lacks in attaining the accurate detection of abnormalities present in the images. In order to mitigate this issue Deep Convolution Neural Network (DCNN) based automatic classification technique is introduced with the classification of eight groups of peripheral blood cells such as basophil, eosinophil, lymphocyte, monocyte, neutrophil, erythroblast, platelet, myocyte, promyocyte and metamyocyte. The proposed DCNN model employs transfer learning approach and additionally it carries three stages such as pre-processing, feature extraction and classification. Initially the pre-processing steps are incorporated to eliminate noisy contents present in the image by using Histogram Equalization (HE). It is enclosed to improve an image contrast. In order to distinguish the dissimilar class and segmentation approach is carried out with the help of Fuzzy C-Means (FCM) model whereas its centroid point optimality method with Slap Swarm based optimization strategy. Moreover some specific set of Gray Level Co-occurrence Matrix (GLCM) features of the segmented images are extracted to augment the performance of proposed detection algorithm. Finally the extracted features are recorded by DCNN and the proposed classifier has the capability to extract their own features. Based on this the diverse set of classes are classified and distinguished from qualitative abnormalities found in the image. [ABSTRACT FROM AUTHOR]

This paper proposes the modified fuzzy min–max neural network (MFMMN) classification model to perform the supervised classification of data. The basic fuzzy min–max neural network (FMMN) can only be applied to the continuous attribute values and cannot handle the discrete values. Also justification of the classification results given by FMMN required to be obtained to make it more applicable to real world applications. These both issues are solved in the proposed MFMMN. In the MFMMN, each hyperbox have min–max values defined in terms of continuous attributes and a set of binary strings defined for discrete attributes. Bitwise ‘ and ’ and ‘ or ’ operators are used to update the discrete values associated with each hyperbox. The trained network is pruned to remove the less useful hyperboxes based on their confidence factor. The proposed model is applied to nine different datasets taken from the University of California, Irvine (UCI) machine learning repository. Finally the case study of a real time weather data is evaluated using MFMMN. The experimental results show that the proposed model has given very good accuracy. In addition to accuracy, the number of hyperboxes obtained after pruning are very less which lead to less number of concise rules and reduced computational complexity. [ABSTRACT FROM AUTHOR]

In this paper we explore a new model focused on integrating two classifiers; Convolutional Neural Network (CNN) and Support Vector Machine (SVM) for offline Arabic handwriting recognition (OAHR) on which the dropout technique was applied. The suggested system altered the trainable classifier of the CNN by the SVM classifier. A convolutional network is beneficial for extracting features information and SVM functions as a recognizer. It was found that this model both automatically extracts features from the raw images and performs classification. Additionally, we protected our model against over-fitting due to the powerful performance of dropout. In this work, the recognition on the handwritten Arabic characters was evaluated; the training and test sets were taken from the HACDB and IFN/ENIT databases. Simulation results proved that the new design based-SVM of the CNN classifier architecture with dropout performs significantly more efficiently than CNN based-SVM model without dropout and the standard CNN classifier. The performance of our model is compared with character recognition accuracies gained from state-of-the-art Arabic Optical Character Recognition, producing favorable results. [ABSTRACT FROM AUTHOR]

Stress is a concerning issue in today's world. Stress in pregnancy harms both the development of children and the health of pregnant women. As a result, assessing the stress levels of working pregnant women is crucial to aid them in developing and growing professionally and personally. In the past, many machine-learning (ML) and deep-learning (DL) algorithms have been made to predict the stress of women. It does, however, have some problems, such as a more complicated design, a high chance of misclassification, a high chance of making mistakes, and less efficiency. With these considerations in mind, our article will use a deep-learning model known as the deep recurrent neural network (DRNN) to predict the stress levels of working pregnant women. Dataset preparation, feature extraction, optimal feature selection, and classification with DRNNs are all included in this framework. Duplicate attributes are removed, and missing values are filled in during the preprocessing of the dataset. [ABSTRACT FROM AUTHOR]

Since the existing music emotion classification researches focus on the single-modal analysis of audio or lyrics, the correlation among models are neglected, which lead to partial information loss. Therefore, a music emotion classification method based on deep learning and improved attention mechanism is proposed. First, the music lyrics features are extracted by Term Frequency-Inverse Document Frequency (TF-IDF) and Word2vec method, and the term frequency weight vector and word vector are obtained. Then, by using the feature extraction ability of Convolutional Neural Network (CNN) and the ability of Long Short-Term Memory (LSTM) network to process the serialized data, and integrating the matching attention mechanism, an emotion analysis model based on CNN-LSTM is constructed. Finally, the output results of the deep neural network and CNN-LSTM model are fused, and the emotion types are obtained by Softmax classifier. The experimental analysis based on the selected data sets shows that the average classification accuracy of the proposed method is 0.848, which is better than the other comparison methods, and the classification efficiency has been greatly improved. [ABSTRACT FROM AUTHOR]

The lack of labels is one of the major challenges in hyperspectral image (HSI) classification. Widely used Deep Learning (DL) models such as convolutional neural networks (CNNs) experience serious performance degradation when training samples are limited. In contrast, graph convolutional networks (GCNs) can simultaneously exploit the insufficient labeled data and massive unlabeled data of HSI in a semisupervised learning fashion. However, in order to reduce computational cost and mitigate noise, existing GCN-based classification methods usually perform superpixel segmentation as a preprocessing step and implement feature extraction as well as node classification on the predefined superpixel graph, where one superpixel might incorporate pixels with different labels. Moreover, the local spectral–spatial information within superpixels is generally ignored. To alleviate these two issues, we propose an end-to-end mixhop superpixel-based GCN (EMS-GCN) framework for HSI classification. Specifically, we first introduce the differentiable superpixel segmentation algorithm to map the pixel representations into a superpixel feature space, which allows refining the superpixel boundary with the training of the network. After that, a superpixel graph is constructed and fed into a novel mixhop superpixel-based GCN, where both the local information within superpixels and long-range information among superpixels are extracted, while the structure of the superpixel graph is updated at the same time. Finally, the enhanced superpixel representations are mapped back into a pixel feature space to conduct pixel-wise classification. Extensive experiments demonstrate the effectiveness of the proposed EMS-GCN method compared with other state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Visual tracking of generic objects is one of the fundamental but challenging problems in computer vision. Here, we propose a novel fully convolutional Siamese network to solve visual tracking by directly predicting the target bounding box in an end-to-end manner. We first reformulate the visual tracking task as two subproblems: a classification problem for pixel category prediction and a regression task for object status estimation at this pixel. With this decomposition, we design a simple yet effective Siamese architecture based classification and regression framework, termed SiamCAR, which consists of two subnetworks: a Siamese subnetwork for feature extraction and a classification-regression subnetwork for direct bounding box prediction. Since the proposed framework is both proposal- and anchor-free, SiamCAR can avoid the tedious hyper-parameter tuning of anchors, considerably simplifying the training. To demonstrate that a much simpler tracking framework can achieve superior tracking results, we conduct extensive experiments and comparisons with state-of-the-art trackers on a few challenging benchmarks. Without bells and whistles, SiamCAR achieves leading performance with a real-time speed. Furthermore, the ablation study validates that the proposed framework is effective with various backbone networks, and can benefit from deeper networks. Code is available at https://github.com/ohhhyeahhh/SiamCAR. [ABSTRACT FROM AUTHOR]

The dual-polarized Sentinel-1 synthetic aperture radar (SAR), which performs C-band SAR imaging day and night regardless of the weather, offers new opportunities for wetland cover monitoring in regions frequently covered by cloud. In this study, a decision tree (DT) classifier was applied to investigate the utility of Sentinel-1 for wetland classification in Poyang Lake, China. Six land cover classes were identified: water body, bare land, aquatic vegetation, cropland, forest, and urban area surrounding the lake. Two types of features were extracted from the dual-polarized SAR data, namely, the backscattering coefficients and the polarimetric decomposition components. Then, the DT classifier was trained and applied with backscattering features, polarimetric features, or a combination of the two features. The overall accuracy of all the classifiers was over 90% for the different feature combinations (98.02, 90.63, and 98.59%) for the classes in the lake area, compared with less than 78% (74.84, 63.01, and 77.29%) when the classes surrounding the lake were also considered, which demonstrates the potential of Sentinel-1 SAR data for wetland monitoring. The accuracy of different feature combinations increased in the following order: polarimetric features < backscattering features < combination of polarimetric and backscattering features. The artificial neural network, naïve Bayes, random forest, and adaptive boost algorithms were compared for the case of using backscattering features, and the adaptive boost algorithm showed lower performance than the other three algorithms. [ABSTRACT FROM AUTHOR]

Uncertainty modeling of renewable energy sources, load demand, electricity price, etc. create a high volume of data in smart grids. Accordingly, in this article, a precise forecasting method based on a deep learning concept with microclustering (MC) task is presented. The MC method is structured based on hybrid unsupervised and supervised clustering tasks by K-means and Gaussian support vector machine, respectively. In the proposed method, the input data sequence is clustered by the MC task, and then the forecasting process is employed. By applying the MC, input data in each hour are categorized into different groups, and a distinctive forecasting unit is allocated to each one. In this way, more clusters and forecasting networks are earmarked for the hours with higher fluctuation rates. The bi-directional long short-term memory (B-LSTM), which is one of the newest recurrent artificial neural networks, is proposed as the forecasting unit. The B-LSTM has bidirectional memory—feedforward and feedback loops—that helps us to investigate both previous and future hidden layers data. The optimal number of clusters in each hour is determined based on the Davies–Bouldin index. To evaluate the performance of the proposed method, in this study, three forecasting tasks including the wind speed, load demand, and electricity price are studied in different periods using the Ontario province, Canada, data set. The results are compared with other benchmarking methods to verify the robustness and effectiveness of the proposed method. In fact, the proposed method, which is equipped with the MC technique and B-LSTM networks, significantly promotes the forecasting results, especially in spike points. [ABSTRACT FROM AUTHOR]

Recently, multiscale spatial features have been widely utilized to improve the hyperspectral image (HSI) classification performance. However, fixed-size neighborhood involving the contextual information probably leads to misclassifications, especially for the boundary pixels. Additionally, it has been demonstrated that deep neural network (DNN) is practical to extract representative features for the classification tasks. Nevertheless, under the condition of high dimensionality versus small sample sizes, DNN tends to be over-fitting and it is generally time-consuming due to the deep-level feature learning process. To alleviate the aforementioned issues, we propose a multiscale context-aware ensemble deep kernel extreme learning machine (MSC-EDKELM) for efficient HSI classification. First, the scene of the HSI data set is over-segmented in multiscale via using the adaptive superpixel segmentation technique. Second, superpixel pattern (SP) and attentional neighboring superpixel pattern (ANSP) are generated by leveraging the superpixel maps, which can automatically comprise local and global contextual information, respectively. Afterward, an ensemble deep kernel extreme learning machine (EDKELM) is presented to investigate the deep-level characteristics in the SP and ANSP. Finally, the category of each pixel is accurately determined by the decision fusion and weighted output layer fusion strategy. Experimental results on four real-world HSI data sets demonstrate that the proposed frameworks outperform some classic and state-of-the-art methods with high computational efficiency, which can be employed to serve real-time applications. [ABSTRACT FROM AUTHOR]

Classification and identification of the materials lying over or beneath the earth’s surface have long been a fundamental but challenging research topic in geoscience and remote sensing (RS), and have garnered a growing concern owing to the recent advancements of deep learning techniques. Although deep networks have been successfully applied in single-modality-dominated classification tasks, yet their performance inevitably meets the bottleneck in complex scenes that need to be finely classified, due to the limitation of information diversity. In this work, we provide a baseline solution to the aforementioned difficulty by developing a general multimodal deep learning (MDL) framework. In particular, we also investigate a special case of multi-modality learning (MML)—cross-modality learning (CML) that exists widely in RS image classification applications. By focusing on “what,” “where,” and “how” to fuse, we show different fusion strategies as well as how to train deep networks and build the network architecture. Specifically, five fusion architectures are introduced and developed, further being unified in our MDL framework. More significantly, our framework is not only limited to pixel-wise classification tasks but also applicable to spatial information modeling with convolutional neural networks (CNNs). To validate the effectiveness and superiority of the MDL framework, extensive experiments related to the settings of MML and CML are conducted on two different multimodal RS data sets. Furthermore, the codes and data sets will be available at https://github.com/danfenghong/IEEE%5fTGRS%5fMDL-RS , contributing to the RS community. [ABSTRACT FROM AUTHOR]

Recently, multimodal cloud samples are utilized to learn completed feature representations for cloud classification. However, the existing methods neglect the related information from other multimodal cloud samples in the learning process, which leads to inadequate learning. In this article, we propose a novel deep model to learn heterogeneous deep features (HDFs) for multimodal ground-based remote sensing cloud classification. Specifically, we first design the convolutional neural network (CNN) extractor to combine the visual information and the multimodal information (MI) to obtain the CNN-based features of multimodal cloud samples. Afterward, we treat the CNN-based features of multimodal cloud samples as the nodes of graph, and utilize the similarity between nodes as the adjacency matrix. We feed the graph and the adjacency matrix into the graph convolutional network (GCN) extractor to obtain the GCN-based features that could capture correlations among multimodal cloud samples using graph convolutional layers. After obtaining CNN-based features and GCN-based features, we concatenate the two kinds of heterogeneous features to represent the multimodal cloud samples. As a result, the concatenated feature contains the visual information, the MI and the related information among multimodal cloud samples. We conduct a series of experiments on the multimodal ground-based cloud database (MGCD), and the experimental results verify that the proposed HDF outperforms state-of-the-art methods. [ABSTRACT FROM AUTHOR]

• This paper imparts a road map for the researchers of biometric recognition system. • Highlights both Unimodal and multimodal biometric system. • Covering diagrammatic view and major characteristics of all the biometric traits. • Focus is given to the feature extraction methods, selection of the classifier. • Datasets used and the accuracy achieved to explore the main ideas of the authors. Biometrics is the branch of science that deals with the identification and verification of an individual based on the physiological and behavioral traits. These traits or identifiers are permanent, unique and can separate one individual from another. Biometric recognition systems integrate complex definitional, technological and operational selection under various contexts. The systems are not going to replace the authentication tools and technologies, but the combination of biometric approaches and authentication methods to help in improving the security aspects of the applications where user cooperation can be inferred. Biometric based recognition methods and tools have become popular for the development of many useful, challenging and widely accepted applications such as security issues, surveillance, forensic investigations, fraudulent technologies, identity access management and access control. These systems also help to identify an individual in group of industrial networks, home/office building and control system. For the successful implementation of the biometric systems, deep artificial neural networks are in great demand. These systems can be built up either on the single modality or multiple modalities. This article explicates the comprehensive and deep survey that compactly and systematically summarizes the literature work done on unimodal and multimodal biometric systems and analyzes the feature extraction techniques, classifiers, datasets, results, efficiency and reliability of the system with high and multi-dimensional perspectives. This article also justifies in detail the classical methods, influential methods and taxonomy based on the biometric attributes. The goal is to aware the researchers of this area regarding various dimensions for the development of biometric systems to enhance the security aspects. The article begins with the fundamentals, types, need of system, challenges, uncertainties, motivations and then to the survey work. The tabular representation prepares for each biometric trait shows the author, year, major findings and results achieved with the synthesis analysis and the evaluation. The article finally ends up with the 3D biometric, a future perspective and concluding remarks. [ABSTRACT FROM AUTHOR]

The identification of DNA sequences as exon and intron is a common problem in genome analysis. The methods used for feature extraction and mapping techniques for the digitization of sequences affect directly the solution of this problem. The existing mapping techniques are not enough to detect coding and noncoding regions in some genomes because the digital representation of each base in a DNA sequence with an integer does not fully reflect the structure of an original DNA sequence. In the entropy-based mapping technique, we could overcome this problem because the technique deepens distinction rates of exon regions, and better reflects the complexity of DNA sequences. Moreover, in the literature, features are extracted by using various statistical techniques. The statistical features to be extracted are chosen by a system designer's experience. The other proposed approach in this study is to carry out the feature extraction using the transfer learning method. Transfer learning and feature extraction are performed automatically by convolutional neural network models as independent of the data set. In this study, we propose a new method to classify DNA sequences as exon and intron using two approaches. In the first approach, the entropy-based numerical technique was used for the numerical representation of DNA sequences. In the second approach, transfer learning was used to extract features. Then, the obtained features were classified by support vector machine and k -nearest neighbors algorithm. As a result of the classification, accurate performance with 97.8% was achieved. The performance of the current method was compared with the other numerical mapping techniques and feature extraction methods. The results showed that the developed method was much more successful than other methods. [ABSTRACT FROM AUTHOR]

Generally, the database is a gathering of data that is arranged for simple storage, retrieval and modernize. This data comprises of numerous structures like text, table, and image, outline and chart and so on. Content-based image retrieval (CBIR) is valuable for calculating the huge amount of image databases and records and for distinguishes retrieving similar images. Rather than text-based searching, CBIR effectively recovers images that are similar like query image. CBIR assumes a significant role in various areas including restorative finding, industry estimation, geographical information satellite frameworks (GIS frameworks), and biometrics; online searching and authentic research, etc. Here different medical database images are considered to the CBIR procedure is done by the proposed strategy. The proposed method considers the input features are shape, texture feature, wavelet feature, and SIFT feature. To retrieve the input image based on the features, the suggested method utilizes artificial neural network (ANN) structure. Back-propagation technique, which is an organizational structure for learning is utilized for training the neural network framework. Trial demonstrates that the proposed work improves the results of the retrieval system. From the outcomes minimizes the image retrieval time and maximum Precision 87.3% in distance based ANN process. [ABSTRACT FROM AUTHOR]

Huge amount of time series data over several domains such as engineering, biomedical and finance, demands the development of efficient methods for the problem of time series classification. The classification of univariate and multivariate time series together using a single architecture is a very difficult task. In this work, a bio-inspired convolutional spiking neural network (CSNN) is proposed for both univariate and multivariate time series. For this, first we develop a simple transformation to convert raw time series sequences into matrices. The CSNN is a three staged framework which include convolutional feature extraction, spike encoding using soft leaky integrate and fire (Soft-LIf) and classification. As spikes generated are differentiable, thus the learning algorithm for CSNN uses error-backpropagation with cyclical learning rates (CLR) and RMSprop optimizer. Additionally, validation based stopping rules are employed to overcome the overfitting which also provides a set of parameters associated with low validation set loss. Thereafter, to demonstrate the accuracy and robustness of proposed CSNN model, we have used University of California (UCR) univariate as well as University of East Anglia (UEA) multivariate datasets to perform the experiments. Moreover, we conduct comparative empirical performance evaluation with benchmark methods and also with recent deep networks proposed for time series classification. Our results reveal that proposed CSNN advances the baseline methods by achieving higher performance accuracy for both univariate and multivariate datasets. It is shown that the CLR with RMSprop optimizer is able to achieve faster convergence, however CLR and adaptive rates are considered competitive to each other. In addition, we also address the optimal model selection and study the effects of different factors on the performance of proposed CSNN. [ABSTRACT FROM AUTHOR]

Deep learning models have been successfully applied to the analysis of various functional MRI data. Convolutional neural networks (CNN), a class of deep neural networks, have been found to excel at extracting local meaningful features based on their shared-weights architecture and space invariance characteristics. In this study, we propose M2D CNN, a novel multichannel 2D CNN model, to classify 3D fMRI data. The model uses sliced 2D fMRI data as input and integrates multichannel information learned from 2D CNN networks. We experimentally compared the proposed M2D CNN against several widely used models including SVM, 1D CNN, 2D CNN, 3D CNN, and 3D separable CNN with respect to their performance in classifying task-based fMRI data. We tested M2D CNN against six models as benchmarks to classify a large number of time-series whole-brain imaging data based on a motor task in the Human Connectome Project (HCP). The results of our experiments demonstrate the following: (i) convolution operations in the CNN models are advantageous for high-dimensional whole-brain imaging data classification, as all CNN models outperform SVM; (ii) 3D CNN models achieve higher accuracy than 2D CNN and 1D CNN model, but 3D CNN models are computationally costly as any extra dimension is added in the input; (iii) the M2D CNN model proposed in this study achieves the highest accuracy and alleviates data overfitting given its smaller number of parameters as compared with 3D CNN. [ABSTRACT FROM AUTHOR]

We proposed a one-dimensional convolutional neural network (CNN) model, which divides heart sound signals into normal and abnormal directly independent of ECG. The deep features of heart sounds were extracted by the denoising autoencoder (DAE) algorithm as the input feature of 1D CNN. The experimental results showed that the model using deep features has stronger anti-interference ability than using mel-frequency cepstral coefficients, and the proposed 1D CNN model has higher classification accuracy precision, higher F-score, and better classification ability than backpropagation neural network (BP) model. In addition, the improved 1D CNN has a classification accuracy rate of 99.01%. [ABSTRACT FROM AUTHOR]

Ground-penetrating radar (GPR) target detection and classification is a challenging task. Here, we consider online dictionary learning (DL) methods to obtain sparse representations (SR) of the GPR data to enhance feature extraction for target classification via support vector machines. Online methods are preferred because traditional batch DL like K-times singular value decomposition (K-SVD) is not scalable to high-dimensional training sets and infeasible for real-time operation. We also develop Drop-Off MINi-batch Online Dictionary Learning (DOMINODL), which exploits the fact that a lot of the training data may be correlated. The DOMINODL algorithm iteratively considers elements of the training set in small batches and drops off samples which become less relevant. For the case of abandoned anti-personnel landmines classification, we compare the performance of K-SVD with three online algorithms: classical online dictionary learning (ODL), its correlation-based variant, and DOMINODL. Our experiments with real data from L-band GPR show that online DL methods reduce learning time by 36%–93% and increase mine detection by 4%–28% over K-SVD. Our DOMINODL is the fastest and retains similar classification performance as the other two online DL approaches. We use a Kolmogorov–Smirnoff test distance and the Dvoretzky–Kiefer–Wolfowitz inequality for the selection of DL input parameters leading to enhanced classification results. To further compare with the state-of-the-art classification approaches, we evaluate a convolutional neural network (CNN) classifier, which performs worse than the proposed approach. Moreover, when the acquired samples are randomly reduced by 25%, 50%, and 75%, sparse decomposition-based classification with DL remains robust while the CNN accuracy is drastically compromised. [ABSTRACT FROM AUTHOR]

This study aims to propose a vision-based method to classify mosquito species. To investigate the efficiency of the method, we compared two different classification methods: The handcraft feature-based conventional method and the convolutional neural network-based deep learning method. For the conventional method, 12 types of features were adopted for handcraft feature extraction, while a support vector machine method was adopted for classification. For the deep learning method, three types of architectures were adopted for classification. We built a mosquito image dataset, which included 14,400 images with three types of mosquito species. The dataset comprised 12,000 images for training, 1500 images for testing, and 900 images for validating. Experimental results revealed that the accuracy of the conventional method using the scale-invariant feature transform algorithm was 82.4% at maximum, whereas the accuracy of the deep learning method was 95.5% in a residual network using data augmentation. From the experimental results, deep learning can be considered to be effective for classifying the mosquito species of the proposed dataset. Furthermore, data augmentation improves the accuracy of mosquito species' classification. [ABSTRACT FROM AUTHOR]

The challenges in hyperspectral image (HSI) classification lie in the existence of noisy spectral information and lack of contextual information among pixels. Considering the three different levels in HSIs, i.e., subpixel, pixel, and superpixel, offer complementary information, we develop a novel HSI feature learning network (HSINet) to learn consistent features by self-supervision for HSI classification. HSINet contains a three-layer deep neural network and a multifeature convolutional neural network. It automatically extracts the features such as spatial, spectral, color, and boundary as well as context information. To boost the performance of self-supervised feature learning with the likelihood maximization, the conditional random field (CRF) framework is embedded into HSINet. The potential terms of unary, pairwise, and higher order in CRF are constructed by the corresponding subpixel, pixel, and superpixel. Furthermore, the feedback information derived from these terms are also fused into the different-level feature learning process, which makes the HSINet-CRF be a trainable end-to-end deep learning model with the back-propagation algorithm. Comprehensive evaluations are performed on three widely used HSI data sets and our method outperforms the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Convolutional neural networks (CNNs) have strong feature extraction capability, which have been used to extract features from the hyperspectral image. Local binary pattern (LBP) is a simple but powerful descriptor for spatial features, which can lessen the workload of CNNs and improve the classification accuracy. In order to make full use of the feature extraction capability of CNNs and the discrimination of LBP features, a novel classification method combining dual-channel CNNs and LBP is proposed. Specifically, a one-dimensional CNN (1D-CNN) is adopted to process original hyperspectral data to extract hierarchical spectral features and another same 1D-CNN is applied to process LBP features to further extract spatial features. Then, the concatenation of two fully connected layers from the two CNNs, which fused features, is fed into a softmax classifier to complete the classification. The experimental results demonstrate that the proposed method can provide 98.52%, 99.54% and 99.54% classification accuracy on the Indian Pines, University of Pavia and Salinas data, respectively. And the proposed method can also obtain good performance even with limited training samples. [ABSTRACT FROM AUTHOR]

Most traditional approaches classify hyperspectral image (HSI) pixels relying only on the spectral values of the input channels. However, the spatial context around a pixel is also very important and can enhance the classification performance. In order to effectively exploit and fuse both the spatial context and spectral structure, we propose a novel two-stream deep architecture for HSI classification. The proposed method consists of a two-stream architecture and a novel fusion scheme. In the two-stream architecture, one stream employs the stacked denoising autoencoder to encode the spectral values of each input pixel, and the other stream takes as input the corresponding image patch and deep convolutional neural networks are employed to process the image patch. In the fusion scheme, the prediction probabilities from two streams are fused by adaptive class-specific weights, which can be obtained by a fully connected layer. Finally, a weight regularizer is added to the loss function to alleviate the overfitting of the class-specific fusion weights. Experimental results on real HSIs demonstrate that the proposed two-stream deep architecture can achieve competitive performance compared with the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Variation in the format and classification requirements for remote sensing data makes establishing a standard remote sensing sample dataset difficult. As a result, few remote sensing deep neural network models have been widely accepted. We propose a hybrid deep neural network model based on a convolutional auto-encoder and a complementary convolutional neural network to solve this problem. The convolutional auto-encoder supports feature extraction and data dimension reduction of remote sensing data. The extracted features are input into the convolutional neural network and subsequently classified. Experimental results show that in the proposed model, the classification accuracy increases from 0.916 to 0.944, compared to a traditional convolutional neural network model; furthermore, the number of training runs is reduced from 40,000 to 22,000, and the number of labelled samples can be reduced by more than half, all while ensuring a classification accuracy of no less than 0.9, which suggests the effectiveness and feasibility of the proposed model. [ABSTRACT FROM AUTHOR]

This article presents the study regarding the problem of dimensionality reduction in training data sets used for classification tasks performed by the probabilistic neural network (PNN). Two methods for this purpose are proposed. The first solution is based on the feature selection approach where a single decision tree and a random forest algorithm are adopted to select data features. The second solution relies on applying the feature extraction procedure which utilizes the principal component analysis algorithm. Depending on the form of the smoothing parameter, different types of PNN models are explored. The prediction ability of PNNs trained on original and reduced data sets is determined with the use of a 10-fold cross validation procedure. [ABSTRACT FROM AUTHOR]

Electromyogram (EMG) signals have been increasingly used for hand and finger gesture recognition. However, most studies have focused on the wrist and whole-hand gestures and not on individual finger (IF) gestures, which are considered more challenging. In this study, we develop EMG-based hand/finger gesture classifiers based on fixed electrode placement using machine learning methods. Ten healthy subjects performed ten hand/finger gestures, including seven IF gestures. EMG signals were measured from three channels, and six time-domain (TD) features were extracted from each channel. A total of 18 features was used to build personalized classifiers for ten gestures with an artificial neural network (ANN), a support vector machine (SVM), a random forest (RF), and a logistic regression (LR). The ANN, SVM, RF, and LR achieved mean accuracies of 0.940, 0.876, 0.831, and 0.539, respectively. One-way analyses of variance and F-tests showed that the ANN achieved the highest mean accuracy and the lowest inter-subject variance in the accuracy, respectively, suggesting that it was the least affected by individual variability in EMG signals. Using only TD features, we achieved a higher ratio of gestures to channels than other similar studies, suggesting that the proposed method can improve the system usability and reduce the computational burden. [ABSTRACT FROM AUTHOR]

In recent years, the development of convolutional neural networks (CNNs) has promoted continuous progress in scene classification of remote sensing images. Compared with natural image datasets, however, the acquisition of remote sensing scene images is more difficult, and consequently the scale of remote sensing image datasets is generally small. In addition, many problems related to small objects and complex backgrounds arise in remote sensing image scenes, presenting great challenges for CNN-based recognition methods. In this article, to improve the feature extraction ability and generalization ability of such models and to enable better use of the information contained in the original remote sensing images, we introduce a multitask learning framework which combines the tasks of self-supervised learning and scene classification. Unlike previous multitask methods, we adopt a new mixup loss strategy to combine the two tasks with dynamic weight. The proposed multitask learning framework empowers a deep neural network to learn more discriminative features without increasing the amounts of parameters. Comprehensive experiments were conducted on four representative remote sensing scene classification datasets. We achieved state-of-the-art performance, with average accuracies of 94.21%, 96.89%, 99.11%, and 98.98% on the NWPU, AID, UC Merced, and WHU-RS19 datasets, respectively. The experimental results and visualizations show that our proposed method can learn more discriminative features and simultaneously encode orientation information while effectively improving the accuracy of remote sensing scene classification. [ABSTRACT FROM AUTHOR]

The discrimination of non-focal class (NFC) and focal class (FC), is vital in localizing the epileptogenic zone (EZ) during neurosurgery. In the conventional diagnosis method, the neurologist has to visually examine the long hour electroencephalogram (EEG) signals, which consumes time and is prone to error. Hence, in this present work, automated diagnosis of FC EEG signals from NFC EEG signals is developed using the Fast Walsh–Hadamard Transform (FWHT) method, entropies, and artificial neural network (ANN). The FWHT analyzes the EEG signals in the frequency domain and decomposes it into the Hadamard coefficients. Five different nonlinear features, namely approximate entropy (ApEn), log-energy entropy (LogEn), fuzzy entropy (FuzzyEn), sample entropy (SampEn), and permutation entropy (PermEn) are extracted from the decomposed Hadamard coefficients. The extracted features detail the nonlinearity in the NFC and the FC EEG signals. The judicious entropy features are supplied to the ANN classifier, with a 10-fold cross-validation method to classify the NFC and FC classes. Two publicly available datasets such as the University of Bonn and Bern-Barcelona dataset are used to evaluate the proposed approach. A maximum sensitivity of 99.70%, the accuracy of 99.50%, and specificity of 99.30% with the 3750 pairs of NFC and FC signal are achieved using the Bern-Barcelona dataset, while the accuracy of 92.80%, the sensitivity of 91%, and specificity of 94.60% is achieved using University of Bonn dataset. Compared to the existing technique, the proposed approach attained a maximum classification performance in both the dataset. [ABSTRACT FROM AUTHOR]

Every pixel in a hyperspectral image contains detailed spectral information in hundreds of narrow bands captured by hyperspectral sensors. Pixel-wise classification of a hyperspectral image is the cornerstone of various hyperspectral applications. Nowadays, deep learning models represented by the convolutional neural network (CNN) provides an ideal solution for feature extraction, and has made remarkable achievements in supervised hyperspectral classification. However, hyperspectral image annotation is time-consuming and laborious, and available training data is usually limited. Due to the "small-sample problem", CNN-based hyperspectral classification is still challenging. Focused on the limited sample-based hyperspectral classification, we designed an 11-layer CNN model called R-HybridSN (Residual-HybridSN) from the perspective of network optimization. With an organic combination of 3D-2D-CNN, residual learning, and depth-separable convolutions, R-HybridSN can better learn deep hierarchical spatial–spectral features with very few training data. The performance of R-HybridSN is evaluated over three public available hyperspectral datasets on different amounts of training samples. Using only 5%, 1%, and 1% labeled data for training in Indian Pines, Salinas, and University of Pavia, respectively, the classification accuracy of R-HybridSN is 96.46%, 98.25%, 96.59%, respectively, which is far better than the contrast models. [ABSTRACT FROM AUTHOR]

Presently, most deaths are caused by heart disease. To overcome this situation, heartbeat sound analysis is a convenient way to diagnose heart disease. Heartbeat sound classification is still a challenging problem in heart sound segmentation and feature extraction. Dataset-B applied in this study that contains three categories Normal, Murmur and Extra-systole heartbeat sound. In the purposed framework, we remove the noise from the heartbeat sound signal by applying the band filter, After that we fixed the size of the sampling rate of each sound signal. Then we applied down-sampling techniques to get more discriminant features and reduce the dimension of the frame rate. However, it does not affect the results and also decreases the computational power and time. Then we applied a purposed model Recurrent Neural Network (RNN) that is based on Long Short-Term Memory (LSTM), Dropout, Dense and Softmax layer. As a result, the purposed method is more competitive compared to other methods. [ABSTRACT FROM AUTHOR]

The use of Convolutional Neural Networks (CNNs) to solve Domain Adaptation (DA) image classification problems in the context of remote sensing has proven to provide good results but at high computational cost. To avoid this problem, a deep learning network for DA in remote sensing hyperspectral images called TCANet is proposed. As a standard CNN, TCANet consists of several stages built based on convolutional filters that operate on patches of the hyperspectral image. Unlike the former, the coefficients of the filter are obtained through Transfer Component Analysis (TCA). This approach has two advantages: firstly, TCANet does not require training based on backpropagation, since TCA is itself a learning method that obtains the filter coefficients directly from the input data. Second, DA is performed on the fly since TCA, in addition to performing dimensional reduction, obtains components that minimize the difference in distributions of data in the different domains corresponding to the source and target images. To build an operating scheme, TCANet includes an initial stage that exploits the spatial information by providing patches around each sample as input data to the network. An output stage performing feature extraction that introduces sufficient invariance and robustness in the final features is also included. Since TCA is sensitive to normalization, to reduce the difference between source and target domains, a previous unsupervised domain shift minimization algorithm consisting of applying conditional correlation alignment (CCA) is conditionally applied. The results of a classification scheme based on CCA and TCANet show that the DA technique proposed outperforms other more complex DA techniques. [ABSTRACT FROM AUTHOR]