Scholarlink

This study leverages Steady-State Visual Evoked Potentials (SSVEP) and Brain-Computer Interface (BCI) technology to classify responses to different visual stimuli among various subjects. Real-time collected SSVEP data was utilized, and based on a sliding window data collection method, a new classification optimization method based on a weighted voting mechanism was designed and proposed. By comparing the accuracy and Information Transfer Rate (ITR) between the traditional and new methods, the new method significantly improved performance metrics, with accuracy increasing from an average of 77.2% to 94.5%, and ITR from 107.61 bits/min to 180.60 bits/min. Additionally, through Monte Carlo simulation experiments, this study explored the optimal weighting ratio, ultimately determining the best weight distribution based on the distribution of experimental accuracy rates.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimizing the Classification of SSVEP Signals in Brain-Computer Interfaces: A Novel Sliding Window Data Segmentation Method Based on Weighted Voting Mechanism

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yuhao Tong,
  • Yang An,
  • Weidong Su

摘要

This study leverages Steady-State Visual Evoked Potentials (SSVEP) and Brain-Computer Interface (BCI) technology to classify responses to different visual stimuli among various subjects. Real-time collected SSVEP data was utilized, and based on a sliding window data collection method, a new classification optimization method based on a weighted voting mechanism was designed and proposed. By comparing the accuracy and Information Transfer Rate (ITR) between the traditional and new methods, the new method significantly improved performance metrics, with accuracy increasing from an average of 77.2% to 94.5%, and ITR from 107.61 bits/min to 180.60 bits/min. Additionally, through Monte Carlo simulation experiments, this study explored the optimal weighting ratio, ultimately determining the best weight distribution based on the distribution of experimental accuracy rates.