The article presents DOO-RE, a dataset aimed at recognizing user activities using machine learning methods and IoT technology. The dataset includes streams from various ambient sensors like Sound and Projector, segmented into activity units. These units are labeled by multiple annotators through a cross-validation annotation process, resulting in nine types of activities. Unlike previous datasets, DOO-RE supports the recognition of both single and group activities in a real meeting room, making it a significant contribution to the field of activity recognition in public spaces.
Publication date: 19 Jan 2024
Project Page: Not provided
Paper: https://arxiv.org/pdf/2401.08962
