When image indexing meets knowledge discovery

In our paper, we deal with the challenge of extending automatically the classic image indexing by visual relationship features. The visual relationship features are discovered automatically from images. They contribute to make more efficient the content-based indexing. More particularly, we develop an advanced content-based indexing articulated around the following notions : - classic indexing, - clustering algorithm, visual feature book and relationship qualification. Keywords Image, Indexing, retrieval, content, similarity, knowledge discovery, relations.

• Type: Conference
• Year: 2000
• Where: KDD
• Authors: Chabane Djeraba
• Keywords: Classic Image Indexing, Content-based Indexing, Data Mining, KDD 2000, Visual Relationship Features