A Generalized Tonic-Clonic Seizure Detection Method in a Bathtub using a Natural Language Verification Step

Abstract

A Generalized Tonic-Clonic Seizure Detection Method in a Bathtub is proposed in this work, targeting mainly epileptic seizures. The main novelty of the presented methodology is found in the use of natural language Human-Computer Interaction, in order to verify the positive detection of such seizure. The proposed system, is a low-cost and low-complexity one, that allows safe operation in an environment with limited access. It overcomes problems presented in related works in the scientific literature and offers a sufficient two-tier verification process, in order to eliminate the false positives. The preliminary results, based on simulation, showed that any incident was successfully detected. To the best of the authors knowledge this is the first time that this methodology is presented, combining smart vibration detection and natural language (NL) recognition to verify/trigger (or not) an alarm of seizure in a bathtub. © 2020 ACM.