Feature based fall detection system for elders using compressed sensing in WVSN

In general there is a steep increase in the number of cases related to elderly people falling down and getting hospitalized since they are living alone. This increases the need for an efficient and low cost surveillance based fall detection system. Wireless video sensor network (WVSN) can be used for such surveillance applications like monitoring elderly people at home, old age homes or hospitals. But there are some limitations in WVSN like memory constraint, low bandwidth and limited battery life. A light weight fall detection algorithm with efficient encoding technique is needed to make WVSN suitable for health care applications. In this paper a simple feature based fall detection system using compressed sensing algorithm is proposed and it is compared with the existing method. This proposed framework shows 82.5% reduction in time and 83.75% reduction in energy compared to raw frame transmission. The average percentage of space saving achieved by this proposed work is 83.81% which shows 30% increase when compared to the existing method.