Spatially aware feature selection and weighting for object retrieval
作者：
Highlights：
• We propose a more accurate similarity measurement for object retrieval.
• Our method improves two features of the BoW model.
• Spatial expansion can incorporate more latent visual words into a query.
• Visual word re-weighting can increase weights of reliable visual words.
• The combination of them can improve both precision and recall.
摘要
•We propose a more accurate similarity measurement for object retrieval.•Our method improves two features of the BoW model.•Spatial expansion can incorporate more latent visual words into a query.•Visual word re-weighting can increase weights of reliable visual words.•The combination of them can improve both precision and recall.
论文关键词：Object retrieval,Bag-of-words,Spatial expansion,Visual word re-weighting
论文评审过程：Received 9 August 2012, Revised 13 June 2013, Accepted 23 September 2013, Available online 29 September 2013.
论文官网地址：https://doi.org/10.1016/j.imavis.2013.09.005
