Browsing by Author "Zabihi, Morteza"

Now showing 1 - 3 of 3

Citation - WoS: 112
Citation - Scopus: 130
Analysis of High-Dimensional Phase Space Via Poincare Section for Patient-Specific Seizure Detection
(IEEE-Inst Electrical Electronics Engineers Inc, 2016) Zabihi, Morteza; Kiranyaz, Serkan; Rad, Ali Bahrami; Katsaggelos, Aggelos K.; Gabbouj, Moncef; İnce, Türker
In this paper, the performance of the phase space representation in interpreting the underlying dynamics of epileptic seizures is investigated and a novel patient-specific seizure detection approach is proposed based on the dynamics of EEG signals. To accomplish this, the trajectories of seizure and nonseizure segments are reconstructed in a high dimensional space using time-delay embedding method. Afterwards, Principal Component Analysis (PCA) was used in order to reduce the dimension of the reconstructed phase spaces. The geometry of the trajectories in the lower dimensions is then characterized using Poincare section and seven features were extracted from the obtained intersection sequence. Once the features are formed, they are fed into a two-layer classification scheme, comprising the Linear Discriminant Analysis (LDA) and Naive Bayesian classifiers. The performance of the proposed method is then evaluated over the CHB-MIT benchmark database and the proposed approach achieved 88.27% sensitivity and 93.21% specificity on average with 25% training data. Finally, we perform comparative performance evaluations against the state-of-the-art methods in this domain which demonstrate the superiority of the proposed method.
Citation - WoS: 64
Citation - Scopus: 76
Automated Patient-Specific Classification of Long-Term Electroencephalography
(Academic Press Inc Elsevier Science, 2014) Kiranyaz, Serkan; İnce, Türker; Zabihi, Morteza; Ince, Dilek
This paper presents a novel systematic approach for patient-specific classification of long-term Electroencephalography (EEG). The goal is to extract the seizure sections with a high accuracy to ease the Neurologist's burden of inspecting such long-term EEG data. We aim to achieve this using the minimum feedback from the Neurologist. To accomplish this, we use the majority of the state-of-the-art features proposed in this domain for evolving a collective network of binary classifiers (CNBC) using multi-dimensional particle swarm optimization (MD PSO). Multiple CNBCs are then used to form a CNBC ensemble (CNBC-E), which aggregates epileptic seizure frames from the classification map of each CNBC in order to maximize the sensitivity rate. Finally, a morphological filter forms the final epileptic segments while filtering out the outliers in the form of classification noise. The proposed system is fully generic, which does not require any a priori information about the patient such as the list of relevant EEG channels. The results of the classification experiments, which are performed over the benchmark CHB-MIT scalp long-term EEG database show that the proposed system can achieve all the aforementioned objectives and exhibits a significantly superior performance compared to several other state-of-the-art methods. Using a limited training dataset that is formed by less than 2 min of seizure and 24 min of non-seizure data on the average taken from the early 25% section of the EEG record of each patient, the proposed system establishes an average sensitivity rate above 89% along with an average specificity rate above 93% over the test set. (C) 2014 Elsevier Inc. All rights reserved.
Citation - WoS: 35
Citation - Scopus: 44
Real-Time Phonocardiogram Anomaly Detection by Adaptive 1d Convolutional Neural Networks
(Elsevier, 2020) Kiranyaz, Serkan; Zabihi, Morteza; Rad, Ali Bahrami; İnce, Türker; Hamila, Ridha; Gabbouj, Moncef
The heart sound signals (Phonocardiogram - PCG) enable the earliest monitoring to detect a potential cardiovascular pathology and have recently become a crucial tool as a diagnostic test in outpatient monitoring to assess heart hemodynamic status. The need for an automated and accurate anomaly detection method for PCG has thus become imminent. To determine the state-of-the-art PCG classification algorithm, 48 international teams competed in the PhysioNet (CinC) Challenge in 2016 over the largest benchmark dataset with 3126 records with the classification outputs, normal (N), abnormal (A) and unsure - too noisy (U). In this study, our aim is to push this frontier further; however, we focus deliberately on the anomaly detection problem while assuming a reasonably high Signal-to-Noise Ratio (SNR) on the records. By using 1D Convolutional Neural Networks trained with a novel data purification approach, we aim to achieve the highest detection performance and real-time processing ability with significantly lower delay and computational complexity. The experimental results over the high-quality subset of the same benchmark dataset show that the proposed approach achieves both objectives. Furthermore, our findings reveal the fact that further improvements indeed require a personalized (patient-specific) approach to avoid major drawbacks of a global PCG classification approach. (C) 2020 The Authors. Published by Elsevier B.V.