In this paper, we propose a camera system which translates Japanese texts in a scene. The system is portable and consists of four components: digital camera, character image extraction process, character recognition process, and translation process. The system extracts character strings from a region which a user specifies, and translates them into English.