Service robots, which co-exist with humans to provide various services, obtain information from sensors placed in an environment and/or sensors mounted on robots. In this paper we newly propose the concept of fourth-person sensing which combines wearable cameras (first-person sensing), sensors mounted on robots (second-person sensing), and distributed sensors in the environment (third-person sensing). The proposed concept takes advantages of all three sensing systems, while removing disadvantage of each of them. The first-person sensing can analyze what a person wearing a camera is doing and details around him/her, so the fourth-person sensing has chance to provide pro-active services, which are triggered by predicted human intention and difficult with the second- and third-person systems, by estimating his intention/behavior. We introduce an example scenario using the fourth-person sensing and show the effectiveness of the proposed concept through experiments.