Main Article Content
Taking care of elderly people is a significant duty. Elderly people are prone to fall-related fatalities including deaths. Fall affects them physically as well as mentally. Even though old adults are monitored manually there are cases when the caretaker is away for a moment and the old adults are at high risk of indoor accidents one such is falling. The time gap between the incident of fall and proper medical treatment is crucial in enhancing the chances of recovery. Thus a digitally assisted approach is required when manual monitoring is unavailable and to ease the same. In this project, deep learning-based digital monitoring is proposed for elderly people in indoor environments to detect falls. The proposed neural network architecture is computationally light, owing to lesser trainable parameters and a less complex pre-processing pipeline involving an optical-flow algorithm. Optical flow has the dual advantage of ensuring the privacy of the individual and requirement of inexpensive hardware which is a common RGBcamera.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.