The concept of smart building includes the optimization of energy usage in a building. One of the possible solutions for this is to adaptively adjust appliances utilization according to activity level in the building. Thus, an intelligent activity estimation system needs to be developed. However, massive annotated dataset is necessary to train the system. Therefore, we propose a system that enables rapid data annotation to collect the massive dataset. The interface of the system is designed to encourage the annotator for a fast annotation process. With the proposed system, an image can be annotated within 4.8 seconds in average.
Conference: 2019 International Conference on Eco Engineering Development, Surakarta, Indonesia
Bens Pardamean, Tjeng Wawan Cenggoro, Bloomest Jansen Chandra, Reza Rahutomo
Published at :