Development of a Life Logging System via Depth Imaging- based Human Activity Recognition for Smart Homes

A life logging system that logs human activities in living environments such as homes and buildings will be an essential component to smart homes or smart environments in the near future. This kind of system provides continuous monitoring and recording of a resident’s daily activities from which one can obtain life or behavioral patterns of life events which could be used for lifecare or healthcare. In this paper, we present a work of developing such a life logging system via depth imaging-based human activity recognition. In our system, a depth imaging device is utilized from which depth silhouettes of human activities are obtained. Then from the silhouettes, human body joints information gets extracted and used in activity recognition, producing life logs. Our system is composed of two key processes: one is training the life logging system, the other running the trained system to record life logs. In the training processes, we have developed tools with the following functions: (i) data collection from a depth camera, (ii) extraction of body joint points from each depth silhouette, (iii) feature generation from the body joint points, finally (iv) training of a activity recognizer (i.e., Hidden Markov Models). With these tools, one can train the life logging system. After training, one can run the trained system which recognizes learned activities and stores life log information in a database for future reference. We believe that our developed life logging system should be beneficial in applications such as smart homes and smart environments. In the near future, we plan to release our system, making it accessible to researchers so that it could be tested and used in research works of smart services for smart homes and more