Atrial Fibrillation Burden Estimation Using Multi-Task Deep Convolutional Neural Network.

Atrial fibrillation (AF) burden is defined as the percentage of time the patient is in AF rhythm during a certain monitoring period. The accurate AF burden estimation from the long-term electrocardiogram (ECG) recordings provides improved prognostic value compared to the traditional binary AF diagnosis (present or absent) using the snapshot ECG. However, the presence of frequent ectopic beats and different noise levels pose a challenge for precise AF burden estimation. For the first time, we hypothesized that a multi-task deep convolutional neural network (MT-DCNN) could accurately estimate the AF burden from the long-term ambulatory ECG recordings. The model consists of AF detection as a primary task and reconstruction of ECG sequence as an auxiliary task using DCNNs. The auxiliary task regularizes the model to learn robust feature representations for efficient AF detection, thereby aiding accurate AF burden estimation. The MT-DCNN is compared with the state-of-the-art rhythm-based, rhythm- and morphology-based approaches. The models are developed and evaluated on a large database of n=84 patients, totaling t=1,900 h of continuous ECG recordings from the LTAF database. The generalization performance is evaluated on three independent datasets (AFDB, NSRDB and LTNSRDB) of n=48 subjects, totaling t=761 h of continuous ECG recordings. On the LTAF test set, the proposed model exhibits a lesser mean absolute AF burden estimation error of 2.8 % over the rhythm-based and the rhythm- and morphology-based approaches. In addition, the MT-DCNN provides better generalization results on independent test datasets and at different noise levels. The results demonstrate that the MT-DCNN can accurately estimate the AF burden from long-term ECG recordings; thus, it has the potential to be used in remote patient monitoring applications for improved diagnosis, phenotyping, and management of AF.

Seizure Types Classification by Generating Input Images With in-Depth Features From Decomposed EEG Signals for Deep Learning Pipeline.

Electroencephalogram (EEG) based seizure types classification has not been addressed well, compared to seizure detection, which is very important for the diagnosis and prognosis of epileptic patients. The minuscule changes reflected in EEG signals among different seizure types make such tasks more challenging. Therefore, in this work, underlying features in EEG have been explored by decomposing signals into multiple subcomponents which have been further used to generate 2D input images for deep learning (DL) pipeline. The Hilbert vibration decomposition (HVD) has been employed for decomposing the EEG signals by preserving phase information. Next, 2D images have been generated considering the first three subcomponents having high energy by involving continuous wavelet transform and converting them into 2D images for DL inputs. For classification, a hybrid DL pipeline has been constructed by combining the convolution neural network (CNN) followed by long short-term memory (LSTM) for efficient extraction of spatial and time sequence information. Experimental validation has been conducted by classifying five types of seizures and seizure-free, collected from the Temple University EEG dataset (TUH v1.5.2). The proposed method has achieved the highest classification accuracy up to 99% along with an F1-score of 99%. Further analysis shows that the HVD-based decomposition and hybrid DL model can efficiently extract in-depth features while classifying different types of seizures. In a comparative study, the proposed idea demonstrates its superiority by displaying the uppermost performance.

Multi-Scale Convolutional Neural Network Ensemble for Multi-Class Arrhythmia Classification.

The automated analysis of electrocardiogram (ECG) signals plays a crucial role in the early diagnosis and management of cardiac arrhythmias. The diverse etiology of arrhythmia and the subtle variations in the pathological ECG characteristics pose challenges in designing reliable automated methods. Existing methods mostly use single deep convolutional neural networks (DCNN) based approaches for arrhythmia classification. Such approaches may not be adequate for effectively representing diverse pathological ECG characteristics. This paper presents a novel way of using an ensemble of multiple DCNN classifiers for effective arrhythmia classification named Deep Multi-Scale Convolutional neural network Ensemble (DMSCE). Specifically, we designed multiple scale-dependent DCNN expert classifiers with different receptive fields to encode the scale-specific pathological ECG characteristics and generate the local predictions. A convolutional gating network is designed to compute the dynamic fusion weights for the experts based on their competencies. These weights are used to aggregate the local predictions and generate final diagnosis decisions. Moreover, a new error function with a correlation penalty is formulated to enable interaction and optimal diversity among experts during the training process. The model is evaluated on the PTBXL-2020 12-lead ECG and the CinC-training2017 single-lead ECG datasets and delivers state-of-the-art performance. Average F1-score of 84.5 % and 88.3 % are obtained for the PTBXL-2020 and the CinC-training2017 datasets, respectively. Impressive performance across various cardiac arrhythmias and the elegant generalization ability for different leads make the method suitable for reliable remote or in-hospital arrhythmia monitoring applications.

Classification of Epileptic Seizure From EEG Signal Based on Hilbert Vibration Decomposition and Deep Learning.

A convolution neural network (CNN) architecture has been designed to classify epileptic seizures based on two-dimensional (2D) images constructed from decomposed mono-components of electroencephalogram (EEG) signals. For the decomposition of EEG, Hilbert vibration decomposition (HVD) has been employed. In this work, four brain rhythms - delta, theta, alpha, and beta have been utilized to obtain the mono-components. Certainly, the data-driven CNN model is most efficient for 2D image processing and recognition. Therefore, 2D images have been generated from one-dimensional (1D) decomposed mono-components by employing continuous wavelet transform (CWT). Next, simultaneous multiple input images in parallel have been directly fed into the CNN pipeline for feature extraction and classification. For evaluation, the EEG dataset provided by the Bonn University has been taken into consideration. Further, a 5-fold cross-validation technique has been applied to obtain generalized and robust classification performance. The average classification accuracy, sensitivity, and specificity reached up to 98.6%, 97.2%, and 100% respectively. The results show that the proposed idea is very much efficient in seizure classification. The proposed idea resourcefully combines the advantages of HVD and CNN to classify epileptic seizures from EEG signal.

Seizure Type Classification Using EEG Based on Gramian Angular Field Transformation and Deep Learning.

classification of seizure types plays a crucial role in diagnosis and prognosis of epileptic patients which has not been addressed properly, while most of the works are surrounded by seizure detection only. However, in recent times, few works have been attempted on the classification of seizure types using deep learning (DL). In this work, a novel approach based on DL has been proposed to classify four types of seizures - complex partial seizure, generalized non-specific seizure, simple partial seizure, tonic-clonic seizure, and seizure-free. Certainly, one of the most efficient classes of DL, convolution neural network (CNN) has achieved exemplary success in the field of image recognition. Therefore, CNN has been employed to perform both automatic feature extraction and classification tasks after generating 2D images from 1D electroencephalogram (EEG) signal by employing an efficient technique, called gramian angular summation field. Next, these images fed into CNN to perform binary and multiclass classification tasks. For experimental evaluation, the Temple University Hospital (TUH, v1.5.2) EEG dataset has been taken into consideration. The proposed method has achieved classification accuracy for binary and multiclass - 3, 4, and 5 up to 96.01%, 89.91%, 84.19%, and 84.20% respectively. The results display the potentiality of the proposed method in seizure type classification.Clinical relevance-gramian angular summation field, seizure types, convolution neural network.

A diagnostic information based framework for super-resolution and quality assessment of retinal OCT images.

High-resolution (HR) retinal optical coherence tomography (OCT) images are preferred by the ophthalmologists to diagnose retinal diseases. These images can be obtained by dense scanning of the target retinal region during acquisition. However, a dense scanning increases the image acquisition time and introduces motion artefacts, which corrupt diagnostic information. Therefore, researchers have a growing interest in developing image processing techniques to recover HR images from low-resolution (LR) OCT images. In this paper, we present an automated super-resolution (SR) scheme using diagnostic information weighted sparse representation framework to reconstruct HR images from LR OCT images. The proposed method performs fast and reliable reconstruction of the LR images. We also propose a 2D- variational mode decomposition (VMD) based OCT diagnostic distortion measure (QOCT) to quantify diagnostic distortion in the reconstructed OCT images. The SR method is evaluated on clinical grade OCT images with the proposed diagnostic distortion measure along with the conventional non-diagnostic measures like the contrast to noise ratio (CNR), the equivalent number of looks (ENL) and the peak signal to noise ratio (PSNR). The results show an average CNR of 4.07, ENL of 58.96 and PSNR of 27.72 dB. An average score of 1.53 is obtained using the proposed diagnostic distortion measure. Experimental results quantify that the proposed QOCT metric can effectively capture diagnostic distortion.

Analyzing the vocal tract characteristics for out-of-breath speech.

In this work, vocal tract characteristic changes under the out-of-breath condition are explored. Speaking under the influence of physical exercise is called out-of-breath speech. The change in breathing pattern results in perceptual changes in the produced sound. For vocal tract, the first four formants show a lowering in their average frequency. The bandwidths BF1 and BF2 widen, whereas the other two get narrowed. The change in bandwidth is small for the last three. For a speaker, the change in frequency and bandwidth may not be uniform across formants. Subband analysis is carried out around formants for comparing the variation of the vocal tract with the source. A vocal tract adaptive empirical wavelet transform is used for extracting formant specific subbands from speech and source. The support vector machine performs the subband-based binary classification between the normal and out-of-breath speech. For all speakers, it shows an F1-score improvement of 4% over speech subbands. Similarly, a performance improvement of 5% can be seen for both male and female speakers. Furthermore, the misclassification amount is less for source compared to speech. These results suggest that physical exercise influences the source more than the vocal tract.

A novel diagnostic information based framework for super-resolution of retinal fundus images.

Advancements in tele-medicine have led to the development of portable and cheap hand-held retinal imaging devices. However, the images obtained from these devices have low resolution (LR) and poor quality that may not be suitable for retinal disease diagnosis. Therefore, this paper proposes a novel framework for the super-resolution (SR) of the LR fundus images. The method takes into consideration the diagnostic information in the fundus images during the SR process. In this work, SR is performed on the zone of interest of the fundus images. Clinical information of the selected zone is captured using the Shannon entropy, the contrast sensitivity index (CSI), the multi-resolution (MR) intra-band energy and the MR inter-band eigen features. The support vector machine (SVM) classifier is used to decide the clinical significance of the zone. Highly accurate learning based SR method or the bicubic interpolation is applied to the selected zone based on the classification output. The method is tested on the Standard Diabetic Retinopathy Database Calibration level 1 (DIARETDB1) and the Digital Retinal Images for Vessel Extraction (DRIVE) databases. Classification accuracy of 85.22% and 85.77% is achieved for the DIARETDB1 and DRIVE databases respectively. The SR performance of the algorithm is quantified in terms of peak signal to noise ratio (PSNR), structural similarity index measure (SSIM) and computational time. The proposed diagnostic information based SR achieves computational time efficiency without compromising with the high resolution (HR) reconstruction accuracy of the fundus image zones.

Multiscale Amplitude Feature and Significance of Enhanced Vocal Tract Information for Emotion Classification.

In this paper, a novel multiscale amplitude feature is proposed using multiresolution analysis (MRA) and the significance of the vocal tract is investigated for emotion classification from the speech signal. MRA decomposes the speech signal into number of sub-band signals. The proposed feature is computed by using sinusoidal model on each sub-band signal. Different emotions have different impacts on the vocal tract. As a result, vocal tract responds in a unique way for each emotion. The vocal tract information is enhanced using pre-emphasis. Therefore, emotion information manifested in the vocal tract can be well exploited. This may help in improving the performance of emotion classification. Emotion recognition is performed using German emotional EMODB database, interactive emotional dyadic motion capture database, simulated stressed speech database, and FAU AIBO database with speech signal and speech with enhanced vocal tract information (SEVTI). The performance of the proposed multiscale amplitude feature is compared with three different types of features: 1) the mel frequency cepstral coefficients; 2) the Teager energy operator (TEO)-based feature (TEO-CB-Auto-Env); and 3) the breathinesss feature. The proposed feature outperforms the other features. In terms of recognition rates, the features derived from the SEVTI signal, give better performance compared to the features derived from the speech signal. Combination of the features with SEVTI signal shows average recognition rate of 86.7% using EMODB database.

Automated detection of heart ailments from 12-lead ECG using complex wavelet sub-band bi-spectrum features.

The complex wavelet sub-band bi-spectrum (CWSB) features are proposed for detection and classification of myocardial infarction (MI), heart muscle disease (HMD) and bundle branch block (BBB) from 12-lead ECG. The dual tree CW transform of 12-lead ECG produces CW coefficients at different sub-bands. The higher-order CW analysis is used for evaluation of CWSB. The mean of the absolute value of CWSB, and the number of negative phase angle and the number of positive phase angle features from the phase of CWSB of 12-lead ECG are evaluated. Extreme learning machine and support vector machine (SVM) classifiers are used to evaluate the performance of CWSB features. Experimental results show that the proposed CWSB features of 12-lead ECG and the SVM classifier are successful for classification of various heart pathologies. The individual accuracy values for MI, HMD and BBB classes are obtained as 98.37, 97.39 and 96.40%, respectively, using SVM classifier and radial basis function kernel function. A comparison has also been made with existing 12-lead ECG-based cardiac disease detection techniques.

Block sparsity-based joint compressed sensing recovery of multi-channel ECG signals.

In recent years, compressed sensing (CS) has emerged as an effective alternative to conventional wavelet based data compression techniques. This is due to its simple and energy-efficient data reduction procedure, which makes it suitable for resource-constrained wireless body area network (WBAN)-enabled electrocardiogram (ECG) telemonitoring applications. Both spatial and temporal correlations exist simultaneously in multi-channel ECG (MECG) signals. Exploitation of both types of correlations is very important in CS-based ECG telemonitoring systems for better performance. However, most of the existing CS-based works exploit either of the correlations, which results in a suboptimal performance. In this work, within a CS framework, the authors propose to exploit both types of correlations simultaneously using a sparse Bayesian learning-based approach. A spatiotemporal sparse model is employed for joint compression/reconstruction of MECG signals. Discrete wavelets transform domain block sparsity of MECG signals is exploited for simultaneous reconstruction of all the channels. Performance evaluations using Physikalisch-Technische Bundesanstalt MECG diagnostic database show a significant gain in the diagnostic reconstruction quality of the MECG signals compared with the state-of-the art techniques at reduced number of measurements. Low measurement requirement may lead to significant savings in the energy-cost of the existing CS-based WBAN systems.

Analysis of physiological signals using state space correlation entropy.

In this letter, the authors propose a new entropy measure for analysis of time series. This measure is termed as the state space correlation entropy (SSCE). The state space reconstruction is used to evaluate the embedding vectors of a time series. The SSCE is computed from the probability of the correlations of the embedding vectors. The performance of SSCE measure is evaluated using both synthetic and real valued signals. The experimental results reveal that, the proposed SSCE measure along with SVM classifier have sensitivity value of 91.60%, which is higher than the performance of both sample entropy and permutation entropy features for detection of shockable ventricular arrhythmia.

An effective fovea detection and automatic assessment of diabetic maculopathy in color fundus images.

Prolonged diabetes causes severe damage to the vision through leakage of blood and blood constituents over the retina. The effect of the leakage becomes more threatening when these abnormalities involve the macula. This condition is known as diabetic maculopathy and it leads to blindness, if not treated in time. Early detection and proper diagnosis can help in preventing this irreversible damage. To achieve this, the possible way is to perform retinal screening at regular intervals. But the ratio of ophthalmologists to patients is very small and the process of evaluation is time consuming. Here, the automatic methods for analyzing retinal/fundus images prove handy and help the ophthalmologists to screen at a faster rate. Motivated from this aspect, an automated method for detection and analysis of diabetic maculopathy is proposed in this work. The method is implemented in two stages. The first stage involves preprocessing required for preparing the image for further analysis. During this stage the input image is enhanced and the optic disc is masked to avoid false detection during bright lesion identification. The second stage is maculopathy detection and its analysis. Here, the retinal lesions including microaneurysms, hemorrhages and exudates are identified by processing the green and hue plane color images. The macula and the fovea locations are determined using intensity property of processed red plane image. Different circular regions are thereafter marked in the neighborhood of the macula. The presence of lesions in these regions is identified to confirm positive maculopathy. Later, the information is used for evaluating its severity. The principal advantage of the proposed algorithm is, utilization of the relation of blood vessels with optic disc and macula, which enhances the detection process. Proper usage of various color plane information sequentially enables the algorithm to perform better. The method is tested on various publicly available databases consisting of both normal and maculopathy images. The algorithm detects fovea with an accuracy of 98.92% when applied on 1374 images. The average specificity and sensitivity of the proposed method for maculopathy detection are obtained as 98.05% and 98.86% respectively.

Exploiting multi-lead electrocardiogram correlations using robust third-order tensor decomposition.

In this Letter, a robust third-order tensor decomposition of multi-lead electrocardiogram (MECG) comprising of 12-leads is proposed to reduce the dimension of the storage data. An order-3 tensor structure is employed to represent the MECG data by rearranging the MECG information in three dimensions. The three-dimensions of the formed tensor represent the number of leads, beats and samples of some fixed ECG duration. Dimension reduction of such an arrangement exploits correlations present among the successive beats (intra-beat and inter-beat) and across the leads (inter-lead). The higher-order singular value decomposition is used to decompose the tensor data. In addition, multiscale analysis has been added for effective care of ECG information. It grossly segments the ECG characteristic waves (P-wave, QRS-complex, ST-segment and T-wave etc.) into different sub-bands. In the meantime, it separates high-frequency noise components into lower-order sub-bands which helps in removing noise from the original data. For evaluation purposes, we have used the publicly available PTB diagnostic database. The proposed method outperforms the existing algorithms where compression ratio is under 10 for MECG data. Results show that the original MECG data volume can be reduced by more than 45 times with acceptable diagnostic distortion level.