A newspaper reading out system with an adaptive information filtering technology to support visually impaired people

In this paper, we present a newspaper reading out system to support visually impaired people. We built an adaptive newspaper reading out system that sorts headlines in order of a user's priority. The system consists of three modules: information filtering module, speech recognition module and text-to-speech synthesis module. These modules present interface modals that are easy for visually impaired people to use. Our primary experiment with five persons suggests that this reading out system with INSOP is useful for visually impaired people to obtain items in which they are interested. This system reduces the time for a user to obtain information that the user needs to 25% compared with the reading out system without INSOP.