Development of a new image retrieval system using vision key obtained from image segmentation

2016 Fiscal Year Final Research Report

• Project/Area Number: 25330133
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Multimedia database
• Research Institution: University of Miyazaki (2015-2016); Osaka City University (2013-2014)
• Principal Investigator: Pyke Tin 宮崎大学, 国際連携センター, 客員教授 (70536961)
• Project Period (FY): 2013-04-01 – 2017-03-31
• Keywords: ビジョンキー / 類似度測度 / リランキング / 画像クラスタリング / 複数特徴画像セグメンテイション / マルコフ定常状態

Outline of Final Research Achievements

In this research, we have introduced a new concept called “vision key” which is obtained from image segmentation in CBIR (Contents Based Image Retrieval), and developed a new method to present it as a query to the database when retrieving related information. We have also developed a flexible similar image retrieval system that can grasp the user's intention depending on TPO (Time, Place and Occasion) and present the retrieval results according to the request, and confirmed the effectiveness through computer experiments.

Free Research Field

大規模画像検索、ビッグデータ、IoT、クラスタリング、データマイニング