Abstract: High retrieval performances in content based vector graphics retrieval system can be attained by adopting relevance feedback algorithms. A new relevance feedback approach based on combining classifiers was proposed, which combines the expected results from the independent nearest neighbor classifiers with only one training sample formed by each positive or negative feedback sample, computes the relevance score of every vector graphics and optimizes the relevance score by introducing the technique called "Bayesian Query Shifting“. The results of the experiment show that the algorithm not only can further improve the precision of the vector graphics retrieval system but also can ensure the recall of the system.

Key words: Combining classifiers,I3ayesian query shifting,Relevance feedback,Vector graphics retrieval