An Approach for Content-based Image Retrieval Using Region Features in Image Database System

摘要

Once an image feature database is created, the number of its fields to store region features is limited. To ensure all region features extracted from each image can be completely stored in the database, it is necessary to control the number of regions segmented according the number of database fields. In this paper, a novel Approach is proposed for region multi-feature extraction, representation and similarity matching. The main benefit of this algorithm is that it can find an appropriate threshold with a binary searching method, and adjust the number of regions segmented automatically.