Content-independent image processing based fall detection

A novel detection concept, content-independent image processing system (CIIP system) is designed in this study. The regular life safety of elders who live alone is concerned. It is desired to detect possible falling behaviors and unconscious situations of elders in the case. We intend to use image processing techniques to detect elders' fall but without the concern of being watched. The idea is to put a very cheap wireless small camera on the user (elder) to recode environment images instead of users' motion images. The image process of this approach needs to find the motion information directly from the image instead of some known contents. Thus, this kind of approach is referred to as the content-independent image processing in our study. The proposed system uses several basic digital image processing skills and voting mechanisms to distinguish fall and unconscious situations from activities of daily living. The experiments conducted in our study show the effectiveness of the proposed approach.