An object based auto annotation image retrieval system

摘要: In this paper, we proposed an auto annotation image retrieval system. our system, was segmented into regions, each of which corresponds to object. The regions identified by region-based segmentation are more consistent with human cognition than those block-based segmentation. According the object's visual features (color and shape), new objects will be map similar clusters obtain its associated semantic concept. The concepts derived training images may not same as real underlying images, because former depend on low-level features. To ameliorate problem, propose a relevance-feedback model learn long-term short-term interests users. experiments show that algorithm outperforms traditional co-occurrence about 19.5%; furthermore, after five times relevance feedback, mean average precision improves from 46% 62.7%.