Accident Finder
To prevent accidents in living space, we need to actively predict accidents and resolve problems leading to accidents. However, predicting accidents exhaustively and efficiently is difficult for most people. Hense, we propose a method that provides users with clues for predicting accidents; users simply take a picture of a real-world object in their living space. In the proposed method, cases of accidents are collected from web pages using the type of the living space and general name of the captured real-world object; these cases yield characteristic words for predicting accidents.