A deep learning based approach for automatic detection of COVID-19 cases using chest X-ray images.

In this global pandemic situation of coronavirus disease (COVID-19), it is of foremost priority to look up efficient and faster diagnosis methods for reducing the transmission rate of the virus severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Recent research has indicated that radio-logical images carry essential information about the COVID-19 virus. Therefore, artificial intelligence (AI) assisted automated detection of lung infections may serve as a potential diagnostic tool. It can be augmented with conventional medical tests for tackling COVID-19. In this paper, we propose a new method for detecting COVID-19 and pneumonia using chest X-ray images. The proposed method can be described as a three-step process. The first step includes the segmentation of the raw X-ray images using the conditional generative adversarial network (C-GAN) for obtaining the lung images. In the second step, we feed the segmented lung images into a novel pipeline combining key points extraction methods and trained deep neural networks (DNN) for extraction of discriminatory features. Several machine learning (ML) models are employed to classify COVID-19, pneumonia, and normal lung images in the final step. A comparative analysis of the classification performance is carried out among the different proposed architectures combining DNNs, key point extraction methods, and ML models. We have achieved the highest testing classification accuracy of 96.6% using the VGG-19 model associated with the binary robust invariant scalable key-points (BRISK) algorithm. The proposed method can be efficiently used for screening of COVID-19 infected patients.

A multi-channel approach for cortical stimulation artefact suppression in depth EEG signals using time-frequency and spatial filtering.

OBJECTIVE: The stereo electroencephalogram (SEEG) recordings are the sate of the art tool used in pre-surgical evaluation of drug-unresponsive epileptic patients. Coupled with SEEG, electrical cortical stimulation (CS) offer a complementary tool to investigate the lesioned/healthy brain regions and to identify the epileptic zones with precision. However, the propagation of this stimulation inside the brain masks the cerebral activity recorded by nearby multi-contact SEEG electrodes. The objective of this paper is to propose a novel filtering approach for suppressing the CS artifact in SEEG signals using time, frequency as well as spatial information. METHODS: The method combines spatial filtering with tunable-Q wavelet transform (TQWT). SEEG signals are spatially filtered to isolate the CS artifacts within a few number of sources/components. The artifacted components are then decomposed into oscillatory background and sharp varying transient signals using tunable-Q wavelet transform (TQWT). The CS artifact is assumed to lie in the transient part of the signal. Using prior known time-frequency information of the CS artifacts, we selectively mask the wavelet coefficients of the transient signal and extract out any remaining significant electrophysiological activity. RESULTS: We have applied our proposed method of CS artifact suppression on simulated and real SEEG signals with convincing performance. The experimental results indicate the effectiveness of the proposed approach. CONCLUSION: The proposed method suppresses CS artifacts without affecting the background SEEG signal. SIGNIFICANCE: The proposed method can be applied for suppressing both low and high frequency CS artifacts and outperforms current methods from the literature.

A Multivariate Approach for Patient-Specific EEG Seizure Detection Using Empirical Wavelet Transform.

OBJECTIVE: This paper investigates the multivariate oscillatory nature of electroencephalogram (EEG) signals in adaptive frequency scales for epileptic seizure detection. METHODS: The empirical wavelet transform (EWT) has been explored for the multivariate signals in order to determine the joint instantaneous amplitudes and frequencies in signal adaptive frequency scales. The proposed multivariate extension of EWT has been studied on multivariate multicomponent synthetic signal, as well as on multivariate EEG signals of Children's Hospital Boston-Massachusetts Institute of Technology (CHB-MIT) scalp EEG database. In a moving-window-based analysis, 2-s-duration multivariate EEG signal epochs containing five automatically selected channels have been decomposed and three features have been extracted from each 1-s part of the 2-s-duration joint instantaneous amplitudes of multivariate EEG signals. The extracted features from each oscillatory level have been processed using a proposed feature processing step and joint features have been computed in order to achieve better discrimination of seizure and seizure-free EEG signal epochs. RESULTS: The proposed detection method has been evaluated over 177 h of EEG records using six classifiers. We have achieved average sensitivity, specificity, and accuracy values as 97.91%, 99.57%, and 99.41%, respectively, using tenfold cross-validation method, which are higher than the compared state of art methods studied on this database. CONCLUSION: Efficient detection of epileptic seizure is achieved when seizure events appear for long duration in hours long EEG recordings. SIGNIFICANCE: The proposed method develops time-frequency plane for multivariate signals and builds patient-specific models for EEG seizure detection.