Multimodal Retrieval using Mutual Information based Textual Query Reformulation

作者：

Highlights：

• Multimodal Retrieval efficiency can be improved by textual query reformulation.

• A graph based keyphrase extraction incorporating correlation of terms is proposed.

• Textual query is expanded with relevant part of narratives and extracted keyphrases.

• Text and image features are combined using a weightlearning model.

• The proposed method improves both image and text retrieval efficiency significantly.

摘要

•Multimodal Retrieval efficiency can be improved by textual query reformulation.•A graph based keyphrase extraction incorporating correlation of terms is proposed.•Textual query is expanded with relevant part of narratives and extracted keyphrases.•Text and image features are combined using a weightlearning model.•The proposed method improves both image and text retrieval efficiency significantly.

论文关键词：Multimodal Retrieval,Query Reformulation,Keyphrase Extraction,Mutual Information,Fisher-LDA

论文评审过程：Received 18 April 2016, Revised 26 September 2016, Accepted 27 September 2016, Available online 5 October 2016, Version of Record 20 October 2016.

论文官网地址：https://doi.org/10.1016/j.eswa.2016.09.039