Activity recognition using imagery for smart home monitoring

In this chapter, we will describe our comprehensive literature survey on using vision technologies for in-home activity monitoring using computer vision techniques, as well as computational intelligence (CI) approaches. Specifically, through our survey of the body of work, we will address the following questions: I.What are the challenges of using standard RGB cameras for activity analysis and how are they solved?II.Why do most existing algorithms perform so poorly in real-world settings?III.Which the design choices should be considered when deciding between wearable cameras or stationary cameras for activity analysis?IV.What does CI bring to the vision world as compared to computer vision techniques in the activity analysis domain? Through our literature survey, as well as based on our own research in both the wearable and non-wearable domain, we share our experiences to enable researchers to make their own design choices as they enter the field of vision-based technologies. We present the hierarchy of the literature survey in Fig. 1.