Abstract

In the paper, we propose a robust fall detection method which combines head tracking and extraction of human shape within a smart home environment equipped with video cameras. A motion history image and an improved code-book background subtraction technique are combined to extract the human shape. An additional motion-based particle filtering head tracker is also used to ensure the robustness of the system. The extracted human shape information and the head tracking results are combined as criteria for judging the occurrence of a fall. The success of the method is confirmed on real video sequences.