Project 63: Cross-Subject Emotion Recognition with Transfer Learning
Contact Information:
Prof. Bao-Liang Lu
Email: bllu@sjtu.edu.cn
Project Description and Objectives:
Emotion plays a critical role in human lives, which affect our behavior and thoughts. As a signal which directly reflects brain activity, electroencephalography (EEG) has been demonstrated to be a reliable and promising indicator of human mental states. However, due to the physical and mental variance of the human brain, traditional machine learning models may fail when training data and testing data are from different subjects. Recently, transfer learning has attracted the attention of many researchers, which has a high potential in developing cross-subject brain-computer interface systems.
In this project, we will investigate cross-subject emotion recognition using EEG and eye movement signals with transfer learning methods. Based on the existing research and dataset, the study aims to classify five basic emotions (happy, sad, neutral, fear, disgust) among different subjects and to discover more facts about human emotions.
Eligibility Requirements:
Interested students should have basic knowledge of machine learning and programming skills in Python
Experience with TensorFlow or PyTorch is preferred.
Main Tasks:
Explore cross-subject emotion recognition with transfer learning methods.
Collect EEG and eye-tracking experimental data if needed.
Finish a research report.
Website: