CYC based query expansion framework for effective image retrieval

We study several semantic concept-based query expansion and re-ranking scheme and compare different ontology-based expansion methods in image search and retrieval. In particular, we exploit the functions of CYC Knowledge Base for concept expansion. Furthermore, we combine CYC with our image retrieval framework - Pixearch to expand the user´s queries and re-rank the searching results. With the visualized baseline results and user´s interactive pruning, the image retrieval precision and recall can yield significantly increase. Preliminary experiments have been able to demonstrate that the proposed retrieval mechanism has the potential to outperform unaided approaches and other query expansion methods.