基于ERP信号的目标检测研究综述
A Review of Target Detection Based on ERP Signal
杨倩倩 1王伟2
作者信息
- 1. 上海理工大学,上海 200093
- 2. 中国人民解放军海军特色医学中心,上海 200093
- 折叠
摘要
目标检测是一个广泛应用于多个领域的研究课题.传统的计算机视觉方法存在复杂度高、泛化能力差、模型对象可迁移性差等缺点,而人脑在目标检测方面具有较强的认知能力和感知能力,仅需要很少的训练样本就可以达到较好的检测效果.基于事件相关电位(Event-Related Potentials,ERP)的脑机接口(Brain-Computer Interface,BCI)是一种全新的提升目标检测性能的方法,它能够将脑内信息与外部环境进行交互.在此基础上,Gerson提出了一种以脑电信号为基础的快速串行视觉呈现(Rapid Serial Visual Presentation,RSVP)范式,包括显示方式、有效视觉范围、目标复杂度、刺激频率和持续时长等.该文还对ERP信号的预处理、特征提取及分类等方法进行了综述,并对该领域的局限性和发展趋势进行了探讨.
Abstract
Target detection is a research topic which is widely used in many fields.The traditional methods of computer vision have some problems,such as high complexity,poor generalization ability and weak mobility of model objects.However,human brain has higher cognitive ability and perception ability in object detection,only a few training samples are needed to achieve better detection results.The Brain-Computer Interface(BCI)based on Event-Related Potentials(ERP)is a new method to improve the performance of target detection,which can interact with information in the brain and the external environment.On this basis,a new pattern of Rapid Serial Visual Presentation(RSVP)based on EEG signals is proposed by Gerson,it includes display mode,effective visual range,target complexity,stimulus frequency and duration.In addition,the methods of ERP preprocessing,feature extraction and classification are reviewed.Finally,the limitation and development trend of this field are discussed.
关键词
事件相关电位/脑机接口/快速串行视觉呈现/目标检测/特征提取Key words
Event-Related Potentials(ERP)/Brain-Computer Interface(BCI)/Rapid Serial Visual Presentation(RSVP)/Target Detection/Feature Extraction引用本文复制引用
基金项目
后勤科研重点项目(BHJ22J024)
后勤科研重点项目(BHJ14C017)
国防科技创新特区项目(19-163-12-ZT-002-002-01)
国防科技创新特区项目(21-163-12-ZT-006-002-14)
出版年
2024
NETL
NSTL
万方数据