Abstract
Lifelog is [the tool to record the life of human]. However, Lifelog has not been spread, because it has search and UI issues. This paper suggests Sound Lifelog that focuses on the sound or voice only. The feature of this paper is to extract 1.contents of the event, 2. emotion from the sound by using speech recognition and emotion recognition. By focusing on sound, this paper aim to have fixed some issues of the current Lifelog.