A dynamically adapted retrieval algorithm for multi-instance image query with heterogeneous features

We proposed an image retrieval algorithm that boosts salient common features from heterogeneous query images and performs the similarity measurement. The retrieval process is carried out in intra and inter feature domains: (1) the intra retrieval utilizes a method of multi-instance query with one feature and; (2) the inter retrieval coordinates the individual intra retrieval results to boost salient common features among query samples. To represent feature saliency, the averaged normalized correlation coefficient is computed in the intra domain and is normalized in the inter domain. The most distinguished point of this algorithm is that it coordinates heterogeneous features for multi-instance retrieval. Up to five different image types could be integrated for retrieval. Simulations demonstrate satisfactory subjective retrieval performances.