Video Sensors

(a) Kinect depth camera, (b) body joints from depth images, (c) tracked faces in RGB image from Kinect.

RGB cameras are usually used for most of the robots; hence they can be generally used for surveillance and behavior detection or prediction.  Moreover, depth cameras (Kinect, RealSense 3D) can also be good candidates to analysis elderly behavior such as fall detection, gesture recognition to follow some commands (such as controlling home appliances or approach near to elderly to do some help etc.), daily activity history creation to help in diet control or medication, abnormal activity pattern analysis to take precaution, facial expression/emotion analysis to propose some actions such as playing music on being sad or checking relatives by connecting via online voice call etc.