A novel ensemble local graph structure based feature extraction network for EEG signal analysis

Unconstrained and constrained face recognition using dense local descriptor with ensemble framework

Face recognition with Symmetric Local Graph Structure (SLGS)

State-of-the-art data analysis tools have to deal with high-dimensional data. Fortunately, the inherent dimensionality of data is often much smaller, as it has an internal structure limiting its degrees of freedom. In most cases, this structure can be approximated using a graph, i.e, a set of nodes connected by edges. Based on this idea, graphs have been largely used in semi-supervised and unsupervised learning. The canonical assumption when incorporating a graph prior is to assume that the signal is smooth with respect to the graph, i.e, that connected pieces of data have similar values. Nevertheless, this hypothesis is insufficient to characterize more complex relations between the data and its structure, the graph. To this end, the field of Graph Signal Processing (GSP) extends the graph smoothness hypothesis to a more general assumption. Its key idea is to think of the signal as a sum of harmonic modes, which are obtained from the eigen-decomposition of the Laplacian, an operator representing the second order derivative. Thus, the essence of GSP resides in the Fourier transform defined by the projection on the Laplacian eigenvectors basis, which allows us to filter graph signals. However, GSP suffers from a counterintuitive irregularity. Contrarily to the classical Fourier basis, the graph Laplacian eigenvectors are not spreading uniformly on the vertex set and can be highly concentrated in a region of the graph. This non-uniformity results in a lack of any intuitive translation operator and consequently poses challenges to generalize some of the classical signal processing theory, e.g., stationarity, uncertainty principles, or sampling. In this thesis, we answer these difficulties by accepting that each node is unique and should be treated specially according to its surrounding structure. This is achieved using an operator called localization that shifts a kernel defined in the spectral domain around a specific node while adapting to the local graph structure. In the first part of the thesis, we focus on harmonic analysis in GSP. We start by making use of the norm of localized kernels to capture the spectral local features of the graph vertices. To illustrate the relevance of this analysis we then introduce structural clustering, an algorithm that groups nodes according to the role they play in the graph. Then, we bound the vertex-frequency concentration of graph filter banks atoms. Because of the irregular structure of graphs, we construct a local uncertainty principle that is, again, driven by the norm of a localized kernel. In the second part of the thesis, we tackle GSP problems from a machine learning perspective. We use the localization operator to extend the notion of stationarity to graph signals. It consists of a graph signal probabilistic framework allowing us to optimally solve inverse problems. Finally, we show how the graph total variation can be used in semi-supervised learning and prove its consistency with respect to the underlying graph manifold.

Automatic voice based disease detection method using one dimensional local binary pattern feature extraction network

In Electroencephalography (EEG) classifiers, processing of EEG signals consists of extracting features from EEG signal and then processing those features to classify the signals. As of today, in most EEG processing systems, the classification part is done on software platform even though the feature extraction is done on hardware. Hence this require an operating system which limits the development of portable medical devices. This paper presents the design and implementation of feature classification in EEG signal processing system on Field Programmable Gate Array (FPGA) hardware platform. The classification is done with two-class Linear Discriminant Analysis (LDA), based on the features extracted using Discrete Wavelet Transform (DWT), for EEG signals obtained through PhysioNet website. The final design has minimum resource utilization, hence is able to implement on Basys 3 Artix-7 FPGA Board platform using SystemVerilog with SO % hardware accuracy. From this result, it is concluded that this design is suitable for developing low cost, marketable products like sleep detectors for automobile divers. Nevertheless, ultimate the goal is to design a full system, which extracts features and classify EEG signals, that can be implemented without using software platform.

Many real-world social networks are decentralized in nature, and the only way to analyze such a network is to collect local views of the social graph from individual participants. Since local views may contain sensitive information, it is often desirable to apply differential privacy in the data collection process, which provides strong and rigorous privacy guarantees. In many practical situations, the local view of a participant contains not only her own connections, but also those of her neighbors, which are private and sensitive for the neighbors, but not directly so for the participant herself. We call such information beyond direct connections an extended local view (ELV), and study two fundamental problems related to ELVs: first, how do we correctly enforce differential privacy for all participants in the presence of ELVs? Second, how can the data collector utilize ELVs to obtain accurate estimates of global graph properties? This paper points out that when collecting ELVs, it is insufficient to apply a straightforward adaptation of local differential privacy (LDP), a commonly used scheme in practice, to protect the privacy of all network participants. The main problem is that an adversarial data collector can accumulate private information on a specific victim from multiple neighbors of the victim; even though the data collected from each neighbor is perturbed under LDP, their aggregate can still violate the victim's privacy. To prevent this attack, we formulate a novel decentralized differential privacy (DDP) scheme, which requires that each participant consider not only her own privacy, but also that of her neighbors involved in her ELV. The stringent privacy requirement of DDP, however, makes it challenging to design an effective mechanism for data collection. Towards this goal, we design a novel multi-phase framework under DDP that enables an analyst to accurately estimate subgraph counts, an important property of social graphs. The main idea is that instead of collecting subgraph counts directly, which would require excessively noise, the analyst first asks individuals about their respective minimum noise scale, which is private information since it depends on the local graph structure, and, thus, must be performed under DDP. For some types of subgraphs, this process is applied recursively, i.e., the analyst asks about the necessary noise to be injected into the private information on the minimum local noise scale required to protect subgraph counts under DDP. As case studies, we instantiate the proposed framework for three common subgraph patterns: triangles, three-hop paths, and k-cliques. Extensive experiments using real data demonstrate that the proposed scheme leads to accurate estimates of global subgraph counts, whereas baseline solutions fail to obtain meaningful result utility.

In general, graph neural networks (GNNs) adopt the message-passing scheme to capture the information of a node (i.e., nodal attributes, and local graph structure) by iteratively transforming, aggregating the features of its neighbors. Nonetheless, recent studies show that the performance of GNNs can be easily hampered by the existence of abnormal or malicious nodes due to the vulnerability of neighborhood aggregation. Thus it is necessary to learn anomaly-resistant GNNs without the prior knowledge of ground-truth anomalies, given the fact that labeling anomalies is costly and requires intensive domain knowledge. Though removing anomalies through unsupervised anomaly detection methods could be a possible solution, it may render unreasonable GNN model performance on target tasks due to the non-differentiable gap between the two learning procedures. In order to keep the effectiveness of GNNs on anomaly-contaminated graphs, in this paper, we propose a new framework named RARE-GNN (Reinforced Anomaly-REsistant Graph Neural Networks) which can detect anomalies from the input graph and learn anomaly-resistant GNNs simultaneously. Extensive experiments on real-world datasets demonstrate the effectiveness of the proposed framework.

Using electroencephalogram (EEG) signals to control intelligent wheelchairs is one of the new control methods for controlling intelligent wheelchairs. The most serious problem in the control process is recognition and classification of EEG signals, especially a variety of EEG signals. To improve the accuracy of EEG classification, improved-common spatial patterns and a deep-belief network algorithm are proposed and used for recognition and classification of EEG signals. A variety of different EEG signals were collected by an Emotiv EPOC+ EEG collector, and the characteristic signals were extracted by improved-common spatial patterns and identified and classified by a deep-belief network algorithm. Simulation results show that the classification accuracy of improved-common spatial patterns and the deep-belief network algorithm is higher than that of the traditional classification method, and a research perspective is provided for the classification of various EEG signals in the future.

Automated EEG-based language detection using directed quantum pattern technique

The classification of electroencephalogram (EEG) signals is a key technique of brain–computer interface (BCI) system. In view of the complexity of EEG signals and the low accuracy in EEG signals recognition, a motor imagery EEG signals classification method with multi-domain fusion based on Dempster–Shafer (D-S) evidence theory is presented in this paper. Firstly, time domain statistics (TDS), autoregressive (AR) model and discrete wavelet transform (DWT) are used to extract features from EEG signals, respectively, and three probabilistic output support vector machine (SVM) classification models are trained based on these three feature sets. Secondly, using the output of each SVM, we construct basic probability assignment (BPA) function and get fusion BPA through D-S evidence theory. Finally, determining the class of test samples based on decision rules. Four databases from BCI competition are employed to evaluate the proposed approach, and the highest classification accuracy reaches 92.83%. Results show that this method acquires higher accuracy and has strong individual adaptability.

DeLiVoTr: Deep and light-weight voxel transformer for 3D object detection

Convolutional neural network (CNN) has shown phenomenal results on several image classification applications. The performance of CNN is improved with deeper architectures. However, increasing depth comes with certain drawbacks like overfitting and need for more computational resources. Thus, it is necessary to choose optimal depth for a network. In this work, the empirical relation between model depth, capacity and complexity of data is estimated using proposed extended VGGNET (EVGGNET). The EVGGNET consists of feature extraction and classification network. The feature extraction network is divided into pre-trained and extended sections which extract the significant features. The classification network uses these features to classify input image into one of the categories. The basic idea behind EVGGNET is to extend the pre-trained feature extraction network by adding convolutional layers which help to establish the model capacity relationship. VGGNET is used as a pre-trained feature extraction network. The experiments are performed on benchmark datasets Caltech 101 and Caltech 256. The results show that the accuracy of EVGGNET is almost exponentially proportional to the complexity of data.

This paper proposes a multi-source data fusion algorithm for rail surface defect detection in both camera-based rail inspection images and ultrasound B-scan images. First, we design a rail surface segmentation algorithm based on image bilateral filtering, Sobel edge detection, and rail surface edge detection to extract the rail surface area. Second, we build a camera and ultrasound data fusion (CUFuse) model for rail surface defect detection, including two main networks: multi-source data feature extraction and multi-scale feature fusion networks. The multi-source data feature extraction network consists of two BoTNet 50 networks as feature extraction networks to extract five stages of features in camera-based images and ultrasound B-scan images. The multi-scale feature fusion network consists of five feature fusion modules to fuse the feature information output by the multi-source data feature extraction network. Finally, we use the CUFuse model to detect the rail surface defect dataset, and output five rail surface state types, including Light, Moderate, Severe, Normal, and Joint. The results show that the accuracy of the CUFuse model is 96.97%, which can accomplish the task of rail surface defect detection on railway sites.

EEG signal classification using universum support vector machine

Epileptic seizure occurs due to neuronal disorder that results in abnormal pattern of brain signal. Electroencephalogram (EEG) signal represents a modest measure of electric flow in a human brain. An EEG is one of the main diagnostic tests for epilepsy. Due to the presence of seizures, normal pattern of brain waves disappears and different other brain waves can be visualized during the recording of EEG. Approximately, 1% of the total population in the world is affected by this disease. This paper is based on a systematic approach for epilepsy detection of human brain by extraction of features and classification of EEG signal. Feature extraction is completed by discrete wavelet transform (DWT) and multilayer perceptron neural network (MLPNN) with deep learning is used for classification. Experimental study of the proposed work is done through Python platform with an encouraging performance.