Object Image Retrieval Using SLIC and Multiple Image Features

カテゴリ: 部門大会

論文No: SS3-6

グループ名: 【C】平成29年電気学会電子・情報・システム部門大会講演論文集

発行日: 2017/09/06

タイトル(英語): Object Image Retrieval Using SLIC and Multiple Image Features

著者名: 和田 成夫(東京電機大学),二敷 拓人(東京電機大学)

著者名(英語): Shigeo Wada|Hiroto Nishiki

キーワード: 画像検索|オブジェクト抽出|画像特徴量|クラスタリング|Image retrieval|Object extraction|Image feature|Clustering

要約(日本語): CBIR and OBIR are powerful tools to search image. Desired images can be obtained from image database without text indexing. We have previously proposed an OBIR using k-means clustering algorithm to reduce influence of background complexity and additive various image noise. In this paper, to improve robustness of the OBIR system, we propose an object image retrieval system using SLIC and multiple image features. Our method can retrieve the same object with a query object image in noisy and complicated background without learning process. To show the effectiveness of our method, experimental retrieval by cross-validation was executed using image database containing both background and object. Precision rate was used to evaluate the performance of image retrieval.

PDFファイルサイズ: 362 Kバイト