Construction Method of a New Type Semantic Network for Image Retrieval
一种新型图像检索语义网络构建方法*

Self-learning semantic network is now a resultful method of achieving high-level image semanteme. For remedying the shortcomings of the existing methods, this paper proposes the construction method of a new type semantic network for the image retrieval. The method supports the image retrieval by combining the three techniques: low-level feature, relevant feedback, and semantic network. The experiment results proved that this method can enhance the system's retrieval efficiency.