ERA is in the process of being migrated to Scholaris, a Canadian shared institutional repository service (https://scholaris.ca). Deposits and changes to existing ERA items and collections are frozen until migration is complete. Please contact erahelp@ualberta.ca for further assistance
- 135 views
- 675 downloads
Human fall detection using multimodal datasets
-
- Author(s) / Creator(s)
-
Falling events are the leading cause of fatal and non-fatal injuries among elderly, and without timely medical care people unable to recover by themselves, it may mean suffering from serious consequences. Designing reliable fall detection systems may help reduce the time for a person to receive an opportune medical care. Falls are events which might be described using multiple sensors, this is because the problem itself involves many features, such as the sudden change in human’s body position which can be sensed by tracking acceleration changes in the three orthogonal directions, or while monitoring activities of daily living of elderly people. This project is aimed to compare the performance of a fall detection system based on a multimodal dataset using common deep learning algorithms, such as Convolutional Neural Networks and Multi-Layer models and feeding them with image data only (RGB and Depth Data), versus when adding signal data from an accelerometer sensor.
-
- Date created
- 2022
-
- Type of Item
- Research Material