Multiple epileptiform waves detection algorithm based on improved VMD and multidimensional feature fusion

Qiwei Cai

doi:10.1016/j.jneumeth.2026.110703

Multiple epileptiform waves detection algorithm based on improved VMD and multidimensional feature fusion

2026-01-28

0

PRE

AI

DOI：10.1016/j.jneumeth.2026.110703

Original

Original request for help

Save

Abstract

En 中文

• An improved VMD with a dynamic tolerance adjustment strategy is adopted to replace traditional band-pass filtering, thereby reducing ringing artifacts and minimizing the misclassification rate of ripples. • Multidimensional features are extracted from low-frequency and high-frequency domains to characterize spikes and ripples, and recursive feature elimination (RFE) is applied to select discriminative features. • A dual-stream 1D CNN integrated with an adaptive scale factor is constructed to extract deep features from raw time-series data, which are then fused with traditional features to enhance the robustness and accuracy of detection.

Multiple epileptiform waves detection algorithm based on improved VMD and multidimensional feature fusion

Abstract

Journal

Organization