EmoWear Data

EmoWear is a wearable physiological and motion dataset for emotion recognition and context awareness. Data is recorded from 49 participants who watched emotionally-eliciting video clips, and reported their emotions via Self Assessment Manikins (SAM). Peripheral tasks of walking, talking, and drinking were also included. The sensor data includes: Acceleration (ACC) and Gyroscope (GYRO) data recorded form the chest, Electrocardiography (ECG), Blood Volume Pulse (BVP), Respiration (RSP), Electrodermal Activity (EDA), and Skin Temperature (SKT). The dataset is provided in three packages: "raw" package with the original collected data, "mat" and "csv" packages with the cleaned and synchronized data in Matlab and Comma Separated Values formats. Sample zip file includes sample data from all three packages of "raw", "mat", and "csv", so that the users can determine the package that meets their needs without a full dataset download. Full data descriptor is published in Nature Scientific Data (pdf). An initial benchmarking of the EmoWear dataset for emotion recognition task using Seismocardiography (SCG) signal is presented at IEEE Transactions on Affective Computing.