eduzhai > Applied Sciences > Engineering >

Epileptic Seizure Prediction A Semi-Dilated Convolutional Neural Network Architecture

  • king
  • (0) Download
  • 20210505
  • Save

... pages left unread,continue reading

Document pages: 8 pages

Abstract: Accurate prediction of epileptic seizures has remained elusive, despite themany advances in machine learning and time-series classification. In this work,we develop a convolutional network module that exploits Electroencephalogram(EEG) scalograms to distinguish between the pre-seizure and normal brainactivities. Since these scalograms have rectangular image shapes with many moretemporal bins than spectral bins, the presented module uses "semi-dilatedconvolutions " to create a proportional non-square receptive field. The proposedsemi-dilated convolutions support exponential expansion of the receptive fieldover the long dimension (image width, i.e. time) while maintaining highresolution over the short dimension (image height, i.e., frequency). Theproposed architecture comprises a set of co-operative semi-dilatedconvolutional blocks, each block has a stack of parallel semi-dilatedconvolutional modules with different dilation rates. Results show that ourproposed solution outperforms the state-of-the-art methods, achieving seizureprediction sensitivity scores of 88.45 and 89.52 for the American EpilepsySociety and Melbourne University EEG datasets, respectively.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×