Emognition system - wearables, physiology, and machine learning for real-life emotion capturing

Despite much progress in laboratory-based emotion and affect recognition, it still remains an unsolved challenge in daily life. One of the obstacles is collecting a large amount of labeled data necessary to train machine learning models. We propose an enhanced Emognition system for collecting emotional samples in everyday life. The system utilizes smartphones to gather self-assessments and contextual data and unobtrusive off-the-shelf wearables to record physiological signals and additional data. We report the results of Emognition system validation after three months of an ongoing large-scale field study LarField23. We also compare the system with its previous version. Finally, we discuss the usefulness of the system in the context of existing emotion psychophysiology datasets and recent advancements in machine learning.