Multimodal concept-dependent active learning for image retrieval

King-Shy Goh, Edward Y. Chang, Wei-Cheng Lai
2004 Proceedings of the 12th annual ACM international conference on Multimedia - MULTIMEDIA '04  
It has been established that active learning is effective for learning complex, subjective query concepts for image retrieval. However, active learning has been applied in a concept independent way, (i.e., the kernel-parameters and the sampling strategy are identically chosen) for learning query concepts of differing complexity. In this work, we first characterize a concept's complexity using three measures: hitrate, isolation and diversity. We then propose a multimodal learning approach that
more » ... es images' semantic labels to guide a concept-dependent, active-learning process. Based on the complexity of a concept, we make intelligent adjustments to the sampling strategy and the sampling pool from which images are to be selected and labeled, to improve concept learnability. Our empirical study on a 300K-image dataset shows that concept-dependent learning is highly effective for image-retrieval accuracy.
doi:10.1145/1027527.1027664 dblp:conf/mm/GohCL04 fatcat:yq5qccle7vf7dnkiz65j2ttx7a