/Microsleep_Detection

a CNN model with transformer reliability

Primary LanguageJupyter Notebook

Microsleep Detection

a CNN model with transformer reliability

Data

Is available at OSF

Description

This a project for comparision between accuracy and loss of different approaches to problem of detecting microsleep from 1st stage of sleep data; somehow conclusion of this research is to show how much microsleep and 1st stage of sleep is similar to eachother.

  • Accuracies different accuracies
  • Losses different losses
  • CNN part of model CNN model
  • Transformer part of model transformer

Advantage

Our best procedure had lower loss than best work done in this problem and got published as an article; this advantage is due to using transformer.