Video analytics for elderly fall and unstable movement detection on an embedded device | |
| Author | Zereen, Aniqua Nusrat |
| Call Number | AIT Diss no.CS-23-01 |
| Subject(s) | Older people--Care--Technological innovations Gerontechnology Video surveillance |
| Note | A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Computer Science |
| Publisher | Asian Institute of Technology |
| Abstract | Real time detection of falls and unstable movement by elderly people is vital to their quality of life and safety. Current research and development in video processing for el der care focuses on detecting falls and detecting and recording elderly peoples’ activities over time. Opportunities for more proactive monitoring of elder’s activity patterns have not been explored. Real time unusual behavior alerts may help prevent accidents rather than speeding up a reaction after the fact. We present an edge processing device inte grated with a cloud computation framework that can be used for activity profiling as well as trigger alerts for falls and unstable motion by elderly people at home. The proposed system uses fixed cameras to track and analyze each visible person in the scene, classi fying their actions into nine ordinary activities, a fall, or unstable movement. An alert notification is sent to caregivers whenever a fall or unstable movement is detected. The major components of the system include an embedded device (NVIDIA JETSON TX2) and cloud-based storage and analysis infrastructure. The system is composed of mod ules for detecting, tracking and recognizing humans, a cascaded hierarchical classifier for nine ordinary activities and falls, and a long short-term memory (LSTM) module to predict unstable movement in video. The system is designed for accuracy, usabil ity, and cost. A prototype system has been subjected to individual module tests along with a field test within a volunteer’s household. It achieved an accuracy of 91.6% for ordinary actions and falls with a recall of 97.02% for unstable motion. The main limi tations are occlusion of one individual by another or by furniture and the relatively high cost of high-quality IP cameras and edge processing devices. Future phases will ex pand deployment to multiple homes with a wide variety of training data to overcome the difference between simulated data and real environment. The main contribution of the dissertation are unique features of the system which include an edge/cloud architec ture with a capable embedded device (NVIDIA Jetson TX2) and a cloud-based storage and analysis infrastructure. The system can recognize nine ordinary human activities, a fall, or unstable movement in a scene. Everyday activities include sitting, standing, walking, bending, clapping, checking time on a watch, talking on the phone, pointing at something, and waving. |
| Year | 2023 |
| Type | Dissertation |
| School | School of Engineering and Technology |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Computer Science (CS) |
| Chairperson(s) | Dailey, Mathew N. |
| Examination Committee(s) | Mongkol Ekpanyapong;Chaklam Silpasuwanchai |
| Scholarship Donor(s) | Bangabandhu Science and Technology Fellowship Trust, Bangladesh;AIT Fellowship |
| Degree | Thesis (Ph.D.) - Asian Institute of Technology, 2023 |