Unsupervised graph-based rank aggregation for improved retrieval
作者：
Highlights：
• A robust unsupervised graph-based rank aggregation function is presented.
• It is targeted for general applicability, such as image, textual, or multimodal retrieval tasks.
• A fusion graph is proposed to gather information and inter-relationship of multiple retrieval results.
• A novel similarity retrieval score is formulated using fusion graphs and minimum common subgraphs.
• The Extensive experimental protocol shows significant gains over state-of-the-art basseline methods.
摘要
•A robust unsupervised graph-based rank aggregation function is presented.•It is targeted for general applicability, such as image, textual, or multimodal retrieval tasks.•A fusion graph is proposed to gather information and inter-relationship of multiple retrieval results.•A novel similarity retrieval score is formulated using fusion graphs and minimum common subgraphs.•The Extensive experimental protocol shows significant gains over state-of-the-art basseline methods.
论文关键词：Rank aggregation,Content-based retrieval,Multimodal retreival,Graph-based fusion
论文评审过程：Received 17 September 2018, Revised 6 February 2019, Accepted 18 March 2019, Available online 21 March 2019, Version of Record 21 March 2019.
论文官网地址：https://doi.org/10.1016/j.ipm.2019.03.008
