Cascade CNN-based model for epileptic seizure diagnostics

Automated seizure detection is a complex task in epilepsy diagnosis, mainly due to the rarity of seizures. Automatic solutions often lead to a high number of false positives. This study proposes a two-stage approach that combines iterative refinement algorithms and convolutional neural networks to minimize false positives in seizure detection task. The method was tested on a real-world EEG dataset using epilepsy-specific metrics and demonstrated a significant reduction in false positives, demonstrating the potential of the method for use in clinical decision-support systems.