Learning a hybrid similarity measure for image retrieval

作者：

Highlights：

• Hybrid similarity learning for preserving both visual and semantic resemblance.

• Using unlabeled images for similarity learning through an asymmetric mechanism.

• Incremental visual similarity learning to address the real-time requirement.

摘要

•Hybrid similarity learning for preserving both visual and semantic resemblance.•Using unlabeled images for similarity learning through an asymmetric mechanism.•Incremental visual similarity learning to address the real-time requirement.

论文关键词：Image retrieval,Relevance feedback,Hybrid similarity measure,Short-term learning,Long-term learning

论文评审过程：Received 11 April 2012, Revised 8 April 2013, Accepted 12 April 2013, Available online 20 April 2013.

论文官网地址：https://doi.org/10.1016/j.patcog.2013.04.008