Automatic Exam Evaluation based on Brain Computer Interface

TL;DR

This paper presents a BCI-based system for automatic exam evaluation that measures student responses via brain signals, achieving 91% accuracy, and examines the impact of noise on system performance.

Contribution

It introduces an online BCI system for educational assessment using P300 signals and demonstrates its effectiveness with high accuracy.

Findings

Achieved 91% accuracy in classifying responses

System works online with real-time brain signal processing

Noise impacts system accuracy, as analyzed in experiments

Abstract

Brain computer interface applications can be used to overcome learning problems, especially student anxiety, lack of focus, and lack of attention. This paper introduces a system based on brain computer interface (BCI) to be used in education to measure intended learning outcomes and measure the impact of noise on the degree of system accuracy. This system works online and is based on recorded brain signal dataset. The system can be considered as a special case of P300 speller accepting only letters from A to D. These are the possible answers to multiple-choice questions MCQ. The teacher makes exams, stores them in an exam database and delivers them to students. Students enroll into the system and record their brain signals. Brain signals go through preprocessing phase in which signals undergo low and high pass filter. Then the signals undergo a subsampling and segmentation. The features obtained are used as inputs to Linear Discriminant Analysis (LDA). Gained accuracy is 91%.