Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
TV VIEWING INTERVAL ESTIMATION FOR PERSONAL PREFERENCE ACQUISITION Hiroaki Tanimoto, Naoko Nitta, and Noboru Babaguchi Graduate School of Engineering, Osaka University 2-1 Yamadaoka, Suita, Osaka 565-0871, Japan tanimoto@nanase.comm.eng.osaka-u.ac.jp, { nitta, babaguchi } @comm.eng.osaka-u.ac.jp ABSTRACT The importance of personalized information services has been increasing. Description of personal preferences needs to be prepared beforehand to realize such services. We propose a system for automatically acquiring personal preferences from TV viewer’s behaviors. Considering “when” a viewer is watch- ing TV is highly related to the viewer’s preferences, we focus on estimating the time interval during which a pre-registered viewer is watching TV. In this paper, we f rstly describe the outline of the personal preference acquisition system, and ad- dress a method for estimating the TV viewing intervals based on the appearance of frontal faces. Experiments resulted in a precision rate of 97.1% and a recall rate of 70.6% on average for TV viewing interval estimation. 1. INTRODUCTION Recently, the importance of personalized information services has been increasing. In order to realize such services, it is essential to prepare description of personal preferences be- forehand. Conventionally, personal preferences have been de- scribed directly by users or acquired by user feedback. How- ever, this can be an excessive burden to users since the data has to be always updated to conform to continuous changes of preferences over time. Therefore, we propose a system for automatically acquiring personal preferences by observ- ing viewer’s behaviors in front of TV[1]. The proposed system records TV viewers by cameras and mi- crophones and recognizes their identi f cation and behaviors by analyzing the recorded video and audio. The TV viewer’s personal preferences are estimated by temporally associating his/her recognized behaviors with the content information of the corresponding video segments. There are some previous works to automatically acquire per- sonal preferences from user’s behaviors. Web browsing his- tories[2], operation records of remote controls[3], and view- ing histories of TV programs[4] have been used as sources to analyze user’s behaviors. Since they have been considering only explicit inputs such as keywords to search when brows- ing Web and fast-forwarding when watching TV, these sys- tems have only enabled us to obtain personal preferences that users have already been aware of. On the other hand, since we consider passive behaviors such as laughing, clapping, and humming to music, even personal preferences which viewers are still unaware of are expected to be obtained. Moreover, traditional personal preference acquisition method from TV viewing histories have only examined viewer’s behaviors for TV program and have not considered the differences of de- gree of interest toward the content. The proposed system tries to acquire more detailed personal preferences by examining
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 2
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 07/21/2011 for the course BUS 10001 taught by Professor All during the Spring '11 term at Shaheed Zulfiqar Ali Bhutto Institute of Science and Technology.

Page1 / 4


This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online