文章摘要

文章信息/Info

Title:: Epilepsy signal recognition based on deep dense convolutional network

Author(s):: CHENG Wei; LONG Jiawei; ZHENG Wei; College of Electronic Information, Jiangsu University of Science and Technology, Zhenjiang 212000, China

Keywords:: Electroencephalogram signal; Epilepsy recognition; Deep learning; Dense network; Convolutional network

摘要:: 癫痫是常见的一种脑部疾病。本研究以德国伯恩大学脑电癫痫信号数据集的预处理版本为样本对象，通过深度卷积神经网络算法DenseNet，实现癫痫脑电信号识别准确率达到96.94%、精确度为97.46%、灵敏度为87.18%、特异度为99.42%和F1分数92.03%的效果。本研究通过Python编码实现密集神经网络，用以进行癫痫脑电信号中的特征识别，以达到识别病灶的目的。本研究具有良好的医学应用前景。

Abstract:: Epilepsy is the most common brain disease.We used the preprocessed version of EEG epileptic signal data set of Bern University in Germany as the sample object, through the deep convolution neural network algorithm DenseNet to realize the recognition accuracy of epileptic EEG signals reached 96.94%, the accuracy was 97.46%, the sensitivity was 87.18%, the specificity was 99.42% and the F1 score was 92.03%. Python coding was used to realize the dense neural network for feature recognition in epileptic EEG signals，and to achieve the purpose of lesion recognition. This study has a good prospect of medical application.