Globally, falls are a major public health problem, and an important cause of morbidity and mortality in the older population. As such, fall detection is one of the most important application areas within the framework of Ambient-Assisted Living (AAL) solutions. Studies report that the majority of falls occur at home, as a person’s living environment is filled with potential hazards, predominantly in the living room and in the bedroom. In addition, recent studies report that fall kinematics varies depending on the weight and size of the falling person, and that most people fall in the evening or during the night. All these features may be captured by RGB-D sensors properly installed in the environment, and detected by suitable processing of the signals generated by the sensors themselves. Fall detection based on RGB-D signal processing has gained momentum in the past years, thanks to the availability of easy-to-use sensors that are able to provide not only raw RGB-D signals but also preprocessed data like joints and skeleton spatial coordinates; additionally, depth signal processing allows to maintain adequate privacy in human monitoring, especially at the levels deemed acceptable by monitored subjects in their own home premises. This chapter will first provide an overview of the RGB-D sensors mostly used in fall detection applications, by discussing their main properties and the modalities by which they have been used and installed. Then, the most relevant signal processing approaches aimed at fall detection will be presented and analyzed, together with an overview of their performances, advantages and limitations, as discussed and presented in the most relevant and up-to-date literature. The aim of the chapter is to provide the reader with a basic understanding of what is reasonably expectable, in terms of detection capability, from RGB-D sensors, applied to fall detection; what are the main depth signal processing approaches according to the sensor usage, and what type of information can be extracted from them.
CITATION STYLE
Spinsante, S. (2019). RGB-D Sensors and Signal Processing for Fall Detection. In Advances in Computer Vision and Pattern Recognition (pp. 309–334). Springer London. https://doi.org/10.1007/978-3-030-28603-3_14
Mendeley helps you to discover research relevant for your work.