Efficient Human Activity Monitoring Using Feature Based Deep Learning Techniques

Abstract

Segmenting behavior-based sensor data and detecting the activity represented by the data are critical elements in any human activity learning applications such as health monitoring, security, and intervention. By recognizing activity transitions, we improve activity recognition. Activity segmentation can be used to increase the performance of activity identification in addition to giving useful activity information. We propose exploiting data obtained from smart homes to identify human activity using machine learning and deep learning methodologies. On data acquired from the five smart homes, we test our suggested segmentation-enhanced activity detection algorithm.