In a context of a growing elderly population in industrialised countries, elderly-care is becoming a majorproblem. As traditional solutions (e.g. retirement homes) are no longer able to satisfy the increasing need,it would be desirable to give the healthiest part of this population a solution for their home care in a secureenvironment.Thus, the concept of smart-home for health, coming from home automation, has emerged in the last twodecades. The general principle of this field is to propose a set of solutions for monitoring compliance, earlydetection of neurodegenerative diseases, social interation assistance or fall and emergency situations detection.Within this framework and to avoid the use of wearable sensors, we proposed a fall detection method basedon optical correlation using video data. Our approach consists of two parts, one for identifying the person onthe picture, the other for head tracking. The detection step is addressed by means of vertical and horizontalcelerity
展开▼
