An Emotion Recognition Dataset Using Millimeter Wave Radar and Physiological Reference Signals
收藏资源简介：
Compared to conventional methods, millimeter-wave (mmWave) radar could obtain vital signs in a non-contact method without privacy concerns. Therefore, we proposed an emotion recognition dataset based on millimeter-wave radar. We used validated stimuli to induce participants' emotions and simultaneously recorded three types of signals: mmWave signals, Photoplethysmogram (PPG) pulse signals, and Galvanic Skin Response (GSR) signals. Participants used the Self-Assessment Manikin (SAM) to provide subjective emotion ratings. We collected signals and emotion rating data from 15 participants and validated the effectiveness of emotion induction and the data quality. The dataset can be used for researches such as mmWave radar based vital sign processing, comparing the performance of different sensing techniques in emotion recognition, and multi-modal fusion emotion recognition, etc. Note: The folder "01_raw_data" contains large files, the files are split into parts. Please download all parts to extract. The dataset involves human physiological data. To obtain the dataset, applicants need to fill out a Data Use Agreement (DUA), providing the research purpose and basic information in the agreement. The DUA can be accessed at: https://github.com/CJ20Project/mmWaveEmoRecDataset-DUA. The signed agreement should be sent to zhouhui@njust.edu.cn. Upon approval, applicants will be granted access to the dataset.
