BS-3-30 On Employing Content-aware Retargeted Image in Automatic Image Annotation(BS-3. Management and Control Technologies for Innovative Networks)
スポンサーリンク
概要
- 論文の詳細を見る
As the number of images on the web is growing exponentially particularly with the help of social networking websites, Automatic Image Annotation (AIA) becomes increasingly important. In this paper, we propose to utilize the retargeted version of the image for image annotation task. We show that by extracting features from this compact and less redundancy version of the image, we could achieve better results. This paper presents some preliminary experiments conducted with the Core15K dataset and some basic image features. The results show a 1 point improvement of the mean average precision (MAP) performance at 95% scale of the original image for one dimension.
- 2012-03-06
著者
-
Sarin Supheakmungkol
Graduate School Of Global Information And Telecommunication Studies (gits) Waseda University
-
Kameyama Wataru
Graduate School Of Global Information And Telecommunication Studies (gits) Waseda University
-
Kameyama Wataru
Graduate School of G lobal Information and Telecommunication Studies, Waseda University
関連論文
- 人間の感情に基づくビデオショットの分類
- D-11-103 Holistic Image Features Extraction for Better Image Annotation
- Special Section on Mobile Multimedia Communications
- D-21 An Adaptive Metadata Filtering Method for Audio-visual Contents Search and Retrieval using User's Viewing History
- Information Distribution Analysis Based on Human's Behavior State Model and the Small-World Network
- D-12-38 RAW RELATIONSHIPS IN AESTHETIC PHOTOS BY OPTICAL FEATURES
- A-15-20 On Consideration of Affective Understanding in Video : A Personalized Approach
- BS-3-30 On Employing Content-aware Retargeted Image in Automatic Image Annotation(BS-3. Management and Control Technologies for Innovative Networks)
- Classification of Video Shots Based on Human Affect