Selective deep convolutional features for image retrieval

Tuan Hoang, Thanh-Toan Do, Dang-Khoa Le Tan, Ngai-Man Cheung

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

34 Citations (Scopus)

Abstract

Convolutional Neural Network (CNN) is a very powerful approach to extract discriminative local descriptors for effective image search. Recent work adopts fine-tuned strategies to further improve the discriminative power of the descriptors. Taking a different approach, in this paper, we propose a novel framework to achieve competitive retrieval performance. Firstly, we propose various masking schemes, namely SIFT-mask, SUM-mask, and MAX-mask, to select a representative subset of local convolutional features and remove a large number of redundant features. We demonstrate that this can effectively address the burstiness issue and improve retrieval accuracy. Secondly, we propose to employ recent embedding and aggregating methods to further enhance feature discriminability. Extensive experiments demonstrate that our proposed framework achieves state-of-the-art retrieval accuracy.

Original languageEnglish
Title of host publicationMM’17 - Proceedings of the 2017 ACM Multimedia Conference
EditorsKuan-Ta Chen, Susanne Boll, Phoebe Chen, Gerald Friedland, Jia Li, Shuicheng Yan
Place of PublicationNew York NY USA
PublisherAssociation for Computing Machinery (ACM)
Pages1600-1608
Number of pages9
ISBN (Electronic)9781450349062
DOIs
Publication statusPublished - 2017
Externally publishedYes
EventACM International Conference on Multimedia 2017 - Mountain View, United States of America
Duration: 23 Oct 201727 Oct 2017
Conference number: 25th
https://dl.acm.org/doi/proceedings/10.1145/3123266

Conference

ConferenceACM International Conference on Multimedia 2017
Abbreviated titleMM 2017
CountryUnited States of America
CityMountain View
Period23/10/1727/10/17
Internet address

Keywords

  • Aggregating
  • Content based image retrieval
  • Deep convolutional features
  • Embedding
  • Unsupervised

Cite this