In the field of automation for accurate diagnosis of Epilepsy (neurological disorder) computer vision and deep learning are linchpin. Epilepsy is diagnosed with the help of Electroencephalogram (EEG) recordings by neurologist or field specialist. EEG interprets recordings in the form of signals. It provides pathological as well as physiological information which is far better than other methods in terms of time taken and technical artifact used for recognizing neurological disorder i.e. epilepsy. EEG signals represents in form of brain waves which shows electrical activity of brain and condition of brain as well. In neuroinformatics, mechanized EEG seizure detection has pre-eminent significance for researchers. Seizures predicted before occurrence definitely results to patient safety with respects to life quality. In this paper LSTM model is used for analysis of EEG signals by adding Gaussian noise on Bonn EEG dataset. Adequacy of the approach is determined on various metrics i.e. accuracy, sensitivity and specificity. The proposed technique has achieved 100% accuracy, 100% sensitivity and 100% specificity only in 30 epochs. The desired results signify that detection approach for seizure possess future for medical convention too.
Epileptic Seizure Detection based on LSTM Model using Noisy EEG Signals
2021-12-02
5259870 byte
Conference paper
Electronic Resource
English
Epileptic Seizure Prediction Method Based on Empirical Mode Decomposition and Kolmogorov Complexity
British Library Online Contents | 2007
|