USF CA Dataset

Welcome! This dataset contains non-posed emotion data sequences for 27 subjects. Each subject id begins with PXXXX. In order to evoke a range of emotions in a lab environment, each participant watches 4 different videos for a total of 3 times. The videos were acquired using sony 4k camcorder and were compressed to a resolution of 640x480. There are also multiple physiological signals acquired during the session using an E4 Empatica watch. The signals include - Heart rate (HR), Temperature (TEMP), 3D accelerometer (ACC) and Electrodermal activity (EDA). Additionally, it also has the facial features extracted from each video using the Openface 2.0 software. More description about this dataset can be found in the paper published at IJCB 2023. Please contact Dr. Shaun Canavan to request the USF CA dataset.